|Publication number||US6629065 B1|
|Application number||US 09/409,569|
|Publication date||Sep 30, 2003|
|Filing date||Sep 30, 1999|
|Priority date||Sep 30, 1998|
|Publication number||09409569, 409569, US 6629065 B1, US 6629065B1, US-B1-6629065, US6629065 B1, US6629065B1|
|Inventors||Rajit Gadh, Tushar H. Dani|
|Original Assignee||Wisconsin Alumni Research Foundation|
|Export Citation||BiBTeX, EndNote, RefMan|
|Patent Citations (7), Non-Patent Citations (31), Referenced by (213), Classifications (12), Legal Events (4)|
|External Links: USPTO, USPTO Assignment, Espacenet|
This application claims priority under 35 USC §119(e) to U.S. Provisional Application No. 60/102,537 filed Sep. 30, 1998, the entirety of which is incorporated by reference herein.
This invention was made with United States government support awarded by the following agencies:
National Science Foundation Grant No.: DMI 9501760
National Science Foundation Grant No.: DMI 9622665
The United States has certain rights in this invention.
This invention relates generally to the early conceptual design of objects using computer-aided design (CAD) systems. More particularly, the invention relates to CAD systems, and methods usable in CAD systems, which allow extremely rapid creation of “rough” or conceptual geometric models of objects without having to precisely describe the specific dimensions, locations, and other characteristics of their geometric subcomponents. The invention further relates to CAD systems, and methods usable in CAD systems, which allow creation and editing of geometric models in such a manner that the created/edited model is output very rapidly after the designer's input is provided (i.e., the final model is produced almost instantaneously after input), so as to enhance the ability to interactively create and edit designs.
The ability to design (i.e., create, edit, visualize, and/or manipulate) two- and three-dimensional models of objects is a fundamental activity in diverse fields such as engineering, architecture, etc. From an engineering perspective, this process is especially important during the conceptual design stage of an object when the final geometric details of the object are unknown. See, e.g., D. L. Jenkins, R. R. Martin, “The Importance of Free-Hand Sketching in Conceptual Design: Automatic Sketch Input,” Proc. of the 1993 ASME Design Technical Conference, Albuquerque, N.Mex., Sep. 19-22, 1993, pp. 115-128. Typical solutions include making use of two-dimensional (2D) sketches to visualize concepts under discussion, or using a Computer-Aided Design (CAD) system to create a prototypical geometric model. Regarding 2D sketching, this suffers from the drawback that it does not readily allow editing and manipulation of the model (the model must be redrawn), and additionally visualization of the 2D-modeled object in three dimensions is difficult. As for CAD, it is a very useful tool, but it is complex and frequently requires detailed user input in order to generate a desired geometric model. For example, when generating a geometric model in conventional CAD systems, significant details must be provided to the CAD system, such as the specific shapes, dimensions, and locations of the geometric subcomponents of the model. Clearly, when working on a final model of an object, the need to enter this degree of detail might be expected if the designer is to obtain the desired amount of precision in the completed model. However, in the early stages of design where the parameters of the object being designed are conceptual and indefinite rather than certain, this need for high human/CAD system interaction is stifling and greatly hinders creativity in the design process. This is especially true where the immediate goal of the designer is to merely capture the rough shape of an object in the CAD system—in essence, to make a rapid “CAD sketch” of the object. In this case, while the designer wishes to take a fast, freewheeling approach to setting forth his or her design concepts, the CAD system's requirements for specificity demand that the designer slow down and provide explicit instructions for all aspects of the object being designed. This informational and time burden increases with the complexity of the CAD system and its human/CAD interface; for example, while modern graphical user interfaces make design of geometric models acceptably rapid in many 2D CAD systems, the ab initio design of objects in 3D CAD systems is still time-consuming and frustrating. The burden grows particularly acute for 3D CAD systems where the human/CAD interface is complex—for example, in Virtual Reality (VR) interfaces wherein the designer works within a Virtual Environment (VE)—owing to the increased processing requirements of such systems.
As a result of the above-noted problems, researchers have attempted to devise methods and apparata which simplify the input process to CAD systems so that a user is not constrained by complex input processes, and is freer to input conceptual designs in a more rapid fashion. As an example, U.S. Pat. No. 5,504,854 to Yamashita illustrates an “Apparatus and Method for Inputting Cell Shape and Position and Inter-Cell Calculation.” This invention creates a series of meta-functions that allow an operator to specify specific shapes and shading using an input device, such as a keyboard or a mouse. When shapes are to be copied or moved, these meta-functions are invoked, thereby reducing the user's burden in specifying the details associated with the shapes to be moved.
To U.S. Pat. No. 5,742,291 to Palm was issued for a “Method and Apparatus for Creation of Three-Dimensional Wire Frames.” This invention addresses rapid creation of representations of three dimensional objects using generically-shaped wire frames which belong to classes of objects. These generic wire frames can then be modified to correspond to a specific shape desired by a user. This avoids the user's need to create individual shapes for subsequent manipulation.
U.S. Pat. No. 5,758,122, to Corda et al. was issued for an “Immersive Visual Programming System.” While this invention is directed to the field of software programming/development, it is nevertheless of interest because it allows the user to select from a set of prefabricated virtual components, thus saving the user the time and effort of creating individual components.
U.S. Pat. No.5,177,689 to Kinasi et al. was issued for a “CAD/CAM Apparatus for Enhancing Operator Entry of Machining Attributes and Geometric Shapes.” This invention relates to the use of simplified forms of input to create numerical control data for machining shapes so that the designer does not need to learn complicated data entry instructions.
U.S. Pat. No. 5,659,493 was issued to Kindena et al. for a “Virtual Machining Techniques for Modifying Computer Models of Parts.” This invention represents three-dimensional objects by a high density data point model. A virtual tool is provided to the user, who manipulates the tool on the surface of the displayed object to modify its shape. Thus, the object can be modified without having to input specific equations or data points associated with the desired modifications.
The above inventions generally use standard man-machine interfaces (e.g., mouse, keyboard, 2D video monitor output) to create designs. However, in more recent years, there has been greater interest in the use of more natural forms of input (such as hand motion input or voice command input), and in Virtual Environments (VE) wherein the user is “immersed” in a 3D design environment, owing to the tremendous potential advantages offered by their use. See, e.g., D. Dietz, “Real Engineering in a Virtual World,” Mechanical Engineering Magazine, vol. 117, no.7, July, 1995, pp. 78-85; T. H. Dani and R. Gadh, “Virtual Reality: A New Tool for the Mechanical Engineer,” Mechanical Engineers Handbook, Myer Kutz, (Ed.), John Wiley & Sons, 1997. As an example, various research projects conducted by the inventors named in this document have investigated inputting shapes into a computer system in a physically intuitive way, using voice, hand motions, and gestures in fashions similar to the way in which one person would communicate with another. By use of position gloves and the like, a CAD system could allow a user to “grab” a design and manipulate it as if holding it in his or her hand, thus providing a more natural and intuitive interaction with the objects that are being designed. This increases design speed and reduces the need to learn complex operating instructions prior to initiating the design process.
Other prior research has also been performed on “natural” forms of user input such as voice commands and hand gestures. See, e.g., D. Weimer, S. K. Ganapathy, “A Synthetic Visual Environment with Hand Gesture and Voice Input”, Proceedings of the ACM Conference on Computer Human Interfaces, May 1989, pp. 235-240; R. A. Bolt, “Put That There: Voice and Gesture at the Graphics Interface,” Computer Graphics, vol. 14, no. 2, pp. 262, 1980. Herranz et al. have developed an interface that utilize two-handed, glove-based input coupled with voice and gaze information to determine the user's intent for different tasks (see, e.g., E. Herranz, R. A. Bolt, “Two-Handed Gesture In Multi-Modal Natural Dialogue”, Proceedings of the ACM Symposium on User Interface Software Technology, Nov. 15-18, 1992, pp. 7-14). Another system that focuses mainly on hand gestures for interaction is the PolyShop environment (D. P. Mapes, J. M. Moshell, “A Two-handed Interface for Object Manipulation in a Virtual Environment,” Presence, vol. 4, no. 4, Fall 1995, pp. 403-416). It allows usage of both hands for various tasks in a VE. Other research has focused on voice and gesture based systems for selection of entities in a geographical map based on spatial relationships; for example, “select the house beside that road,” etc. (Y. A. Tijerno, K. Mochizuiki, F. Kishino, “Interactive 3-D Computer Graphics Driven Through Verbal Instructions: Previous and Current Activities at ATR,” Computer & Graphics, vol. 18, no.5, pp.621-631, 1994.) With the exception of the research reported by Weimer et al., the foregoing systems were not specifically intended for CAD, but are reported here for completeness.
Other research has addressed input devices suitable for use in VE environments. As an example, VE pointers/wands have been developed which allow users to point towards displayed objects to select or otherwise manipulate them. Notable among these are the “BeamCursor” (see, e.g., T. Yoshimura, Y. Nakamura, M. Sugiura, “3D Direct Manipulation Interface: Development of Zashiki-Warashu System,” Computers & Graphics, vol. 18, no. 2, pp. 201-207, 1992) and the “LaserGun” (see, e.g., J. Liang, “JDCAD: A Highly Interactive 3D Modeling System,” Computer Graphics, v. 18, no. 4, 1994, pp. 499-506).
Other research has extended beyond modes of user input and has specifically addressed the use of VE in CAD systems. Some researchers have approached the use of VE in CAD systems by creating models using conventional CAD systems and then “importing” the models (after an appropriate translation process) into a VE for display. This approach faces several difficulties:
(1) The model must still be first defined in a CAD system using precise shapes, dimensions, locations, etc. (which, as noted above, are typically not well-defined at the concept stage). Thus, the aforementioned difficulties with specificity of input and time burdens are still present.
(2) During the process of translation, topological relationships between entities in the model may be lost. As a consequence, editing the model requires a return to the CAD environment so that the desired changes can be made, and the model must then be “re-imported” into the VE for verification.
Examples of such approaches for visualization are the Interactive Visualizer developed at the Graphics, Visualization & Usability Center of the Georgia Institute of Technology (Atlanta, Ga., USA, WWW URL http://www. cc. gatech. edu/gvu/virtual/index_main. html, and the VENUS prototyping project being developed at CERN, the European Center for Nuclear Research (details available at WWW URL: http://sgvenus. cern. ch/VENUS/vr_project. html).
A more useful application of Virtual Environments (VE) that goes beyond these “visualization only” systems is the interactive design of shapes in a VE. Such a system gives the engineer a better design tool that is especially useful during the conceptual design stage. As an example, the aforementioned research by Weimer et al. describes a CAD interface using a VPL DataGlove™ (for capturing hand gestures) and spoken commands (with a vocabulary of about 40 phrases) for creating CAD models. Modeling tools are provided for both free-form shape design and parametric shape design (i.e., design using predefined shape features). The parametric design tools allow object/feature selection and manipulation, an additionally incorporate geometric constraints for object placement. The free-form design tools allow “sculpting” of surfaces that are represented as uniform b-splines and which can be reshaped by perturbing the control points.
Another system is 3-Draw which allows the design of wire-frame like sketches of 3D free-form curve-based geometry (see, e.g., E. Sachs, A. Roberts, D. Stoops, “3-Draw: A Tool for Designing 3D Shapes,” IEEE Computer Graphics & Applications, Vol. 11, no. 11, pp. 18-26, 1991). In this system, the designer uses a pair of 6DOF (six degree-of-freedom) tracking devices (one in each hand). One hand is used to hold the object being sketched, and the other hand holds a stylus which serves as the 3D sketching tool. Curves are represented by first-degree cardinal splines interpolated from points in space obtained via the Polhemus™ tracker. The developers of this system also describe user interaction methods for curve selection, creation and editing. Design tools for curve-shaping included methods based on constrained curve manipulation and methods based on direct manipulation of control points.
While the previous two CAD systems were primarily directed towards creating free-form surface and curve geometry respectively, a system specifically oriented towards the design of engineering components is the aforementioned JDCAD system of Liang (for which a citation was provided earlier). This system aims to provide an intuitive, easy-to-use modeling environment for early stage design. Its researchers describe interaction techniques for navigation, object selection and editing based on 3D menus and icons which are selected via a 6DOF device. The modeling operations are based on creation of primitives via 3D rubber-banding techniques and the use of CSG (Constructive Solid Geometry) operations to create more complex geometry. Region-based reshaping, which involves changing primitive dimensions via “handles” situated thereon, is used for editing. Rotational and translational constraints combined with a 3D snapping feature act as design aids during model creation.
Trika, Banedee and Kashyap (“Virtual Reality Interfaces for Feature-Based Computer-Aided Design Systems”, Computer-Aided Design, vol. 29, no. 8, 1997, pp. 565-574) investigated use of a wand interface (a three dimensional mouse) and a speech input system to simplify the input of design features. This research addressed the issue of the complexity of feature-based and constraint-based design within CAD systems. Feature-based design starts with simple shapes which are then modified by attaching features, such as holes, slots, etc. to modify the simple shapes into the one desired. Constraint-based design involves specifying existing features and the relationship of those features to one another.
Fa et al. propose an approach to 3D design using 3D input devices such as Spaceball™ and Dataglove™ input devices (see, e.g., M. Fa, T. Fernando, and Dew, P. M., “Interactive Constraint-based Solid Modeling using Allowable Motion”, Proc. of the 2nd Symposium on Solid Modeling and Applications, 1993, pp.242-252). Although the proposed system is not strictly a VR-CAD environment, the system integrates so-called direct manipulation techniques with a constraint-based assembly design environment. Constraints are automatically recognized based on user interaction with the 3D devices and are used to control the allowable motion of the individual components in a design assembly. A CSG (Constructive Solid Geometry) graph structure is used to store the design, and each component therein is represented by its boundary representation.
The foregoing systems illustrate that current approaches for VR-based CAD have essentially been based on adapting existing CAD user interfaces within VEs. While this approach has the advantage of being straightforward, it (1) does not exploit the complete potential of VEs, and (2) does not adequately address the development of design tools and representations appropriate for the VR environment. A careful analysis of the differences between conventional CAD and VR-CAD systems shows that their requirements are quite different. In large part, this arises from the differences between their user interfaces, and therefore it is helpful to briefly review these interfaces in greater detail.
Most conventional CAD systems are utilize Windows-Icons-Menu-Pointer (WIMP)-based interfaces to an underlying geometric modeler (FIG. 1). In their most typical form, these are text- and mouse-based interfaces, although in some cases 6DOF input devices such as a Spaceballf and dials may also be used. The designer uses these devices to manipulate designs via design editing tools, e.g., by “sketching” on 2D planes, creating extrusions, etc. These devices are integrated within the CAD environment via an “event driven” approach whereby the designer performs a certain action, and the system responds by performing the necessary computations and displaying the results on the graphical monitor. Although it is desirable that the geometric modeler compute the results and display the image as fast as possible, i.e., with minimum delay, the lack of real-time response is not critical in most conventional CAD applications.
However, while the lack of real-time response is not critical in a conventional CAD system, such response is of great importance in a VR-CAD system. Quite simply, if the design does not respond to the user's inputs almost immediately upon their occurrence, the interface is no longer a “true” VE environment: it does not simulate real-world actions and experiences, and since the motion, voice, etc. no longer have a logical/intuitive link to their outputs, they can be more hindering than helpful. Other differences between CAD and VR-CAD systems center on the multimodality, immersivity, and interactivity of a VE-based application (see, e.g., N. I. Durlach, and A. S. Mavor (Eds), “Part I: Basic Concepts and Terminology,” Virtual Reality: Scientific and Technological Challenges, National Academy Press, Washington D.C., 1995). Each of these concepts will now be briefly discussed.
Multimodality involves providing more than one way for the designer to interact with the model, for example, allowing the user to select, grasp and move entities by both keyboard text input and motion input via a motion sensor. Given that such multiple modalities are available to the designer in a VR-CAD system, it is important to provide design-editing tools that optimize these capabilities both individually and collectively (i.e., both taken alone and when interacting) for shape design.
Immersion is the ability of a VE to impart the designer with the feeling of “presence” in the environment being simulated. An implicit assumption is that the better the immersion (the more natural the behavior of the VE), the more productive the designer would be in accomplishing design objectives. Exemplary immersion methods include providing stereoscopic (3D) images and integrating other feedback methods (such as auditory feedback) in the VR-CAD environment.
Interactivity is primarily involved with the aforementioned need for real-time response; as noted, a key difference between CAD and VR-CAD systems is that interactivity assumes great importance in a VR-based system. In the context of a VE, two measures for interactivity are (1) the frame-rate, and (2) the delay. (See, e.g., N. I. Durlach and A. S. Mavor (Eds), “Part II: Computer Hardware & Software for the Generation of a VE,” Virtual Reality: Scientific and Technological Challenges, National Academy Press, Washington D.C., 1995). The frame-rate refers to the number of graphical scene updates per second, whereas the delay is the lag perceived by the designer between performing a certain action in the VE and perceiving a response on the graphical display. A minimum rate of 10 frames/second is considered acceptable for a VE application, and a 0.1 second delay is identified as being the maximum allowable in a VE. (See, e.g., P. Zhigeng, Z. Mingmin, Z. Wenting, S. Jiaoying, “Time-Critical Computing in Virtual Environments,” Proc. of the 4th International Conference on Computer-Aided Design and Computer Graphics, Oct. 23-25, 1995, Wuhan, China, pp. 712-717.) Both the frame-rate and delay can have significant impact on the interactivity, and hence the effectiveness, of a VE system. The aim in any VE system is to minimize the delay and maximize the frame-rate; however, this requires significant computational resources that remain largely unavailable today.
In large part, difficulties with interactivity arise from the various computations that must be performed in a VR-CAD application, such as intersection detection, rendering, and modeling. Intersection detection is required to detect interference of the user's hand with a virtual object, or collision between two virtual objects, and is important when performing selection of objects and then manipulation/editing. Additionally, intersection detection is used to provide feedback to the user to indicate that contact has been made with an object, as by visual feedback (highlighting) or auditory feedback.
Rendering involves creation of the display of the VE, including any geometry created by the designer, as well as a graphical representation of the user in the VE (e.g., a display of the designer's hand). Current research in rendering has focused on achieving acceptable performance by trading off the accuracy of the various software algorithms used (whether they be for intersection detection, rendering, or modeling) for speed. An example of speeding up graphical rendering by reducing the level of detail of images displayed is given in Zhigeng et al. (P. Zhigeng, Z. Mingmin, Z. Wenting, S. Jiaoying, “Time-Critical Computing in Virtual Environments,” Proc. of the 4th International Conference on Computer-Aided Design and Computer Graphics, Oct. 23-25, 1995, Wuhan, China, pp. 712-717). Another example is provided by Hubbard, wherein objects in the VE are modeled as spheres for approximate collision detection (P. M. Hubbard, “Approximating Polyhedra with Spheres for Time-Critical Collision Detection,” ACM Transactions on Graphics, v. 15, no. 3, Jul. 1996, pp. 179-201). Representing the underlying complex polyhedral geometry of the objects by such spheres enables intersection detection to be completed in a computationally efficient manner.
Modeling relates to the way in which the geometric model is represented and processed within the VR-CAD application. While significant research has been reported on intersection detection and rendering methods for VE applications, modeling has not been researched as extensively. Since the modeling techniques used will have a significant effect on the performance of a VR-CAD system, it will be helpful to briefly review preexisting research on modeling.
Most basic geometric modeling operations in CAD systems utilize a Constructive Solid Geometry (CSG) approach, wherein objects are modeled by primitive geometric elements combined via basic geometric operations such as union, intersection, and negation operations. Since some of these operations are computationally burdensome, some prior researchers have reported attempts to enhance speed by use of specialized or modified operations. As an example, Requicha describes a “restricted CSG” approach in which the only allowable operation is a “glue” operator that can be applied to solids with disjoint interiors. (See, e.g., A. G. Requicha, “Representation for Rigid Solids: Theory Methods & Systems”, Computer Survey, v12, no.4, Dec. 1990, pp 438-464.) Wyvill and Samet describe an octree-based representation (a “DAG-CSG Tree”) in which objects are modeled in geometric space as a combination (via addition and subtraction) of previously defined primitive objects; in conceptual space, a Directed Acyclic Graph (DAG) is used to represent the solids, with the primitives at the nodes represented by octree data structures. (See, e.g., G. Wyvill, T. L. Kunii, “A Functional Model for Constructive Solid Geometry”, The Visual Computer, v. 1, 1985, pp. 3-14.; H. Samet, “Chapter 5: Volume Data, 5.5 Constructive Solid Geometry”, In The Design and Analysis of Spatial Data Structures, Addison Wesley Publishing Company Inc. 1990.) The DAG results in a tree with nodes that contain a flag indicating whether the primitive was added or subtracted, and matrices describe the relative spatial location and shape of the node. General background information regarding DAGs can be found in, e.g., A. V. Aho, J. E. Hopcroft and J. D. Ullman, Data Structures and Algorithms, Addison Wesley Publishing Company, 1I 1983.
Tree and graph-based structures in conceptual space have been used for assembly modeling wherein the nodes of the graph represent individual parts in the assembly and the arcs of the graph represent affine transformations and other information about the mating parts. The Virtual Link concept proposed by Lee et al. has each arc of an assembly graph store complete information about the type of attachment between the parts (rigid, constrained, etc.), as well as information about the mating surfaces of the parts. (See, e.g., K. Lee, D. C. Gossard, “A Hierarchical Data Structure for Representing Assemblies: Part 1,” CAD Journal, v. 17, no. 1, January/February 1985.) The Relationship Graph proposed by Fa et al. (for which a citation was previously given) maintains geometric relationships between parts and assemblies that are represented as CSG (Constructive Solid Geometry) trees. The authors claim that this representation can be used for interactive design by use of 3D hand input devices, but it is unclear as to how the representation will achieve interactive performance suitable for a VE-based system unless a computer system with high processing speed is available.
Other researchers have more specifically addressed modeling methods suitable for VE applications. Virtual Reality Modeling Language (VRML) uses the Scene-Graph representation, which organizes objects into a graph hierarchy in which the leaves of the graphs are the triangular-faceted objects, and the nodes of the graphs are transformations or other such operations. The hierarchy facilitates grouping of objects and allows for optimization of some aspects of collision detection and Scene-Graphs are popular because they provide a representation suitable for real-time displays of predefined static 3D scenes. However, they are not suitable for a VE wherein the geometry may be constantly changing as the user interactively modifies it.
Otherwise, triangular faceted representations of geometric models are typically the representations of choice in VE applications because most current graphics libraries are optimized for them, and they can be accommodated in almost all shape modeling software. However, such a representation has several disadvantages. First, it requires a large amount of memory to store a model, since a model could potentially be composed of millions of triangles. Second, performing editing operations on polygonal models is complex (see, e.g., J. B. Allan, B. Wyvill, I. H. Witten, “A Methodology for Direct Manipulation of Polygon Meshes,” New Advances in Computer Graphics: Proceedings of CD International '89, pp. 451-469). A typical approach to overcome these difficulties is to use a dual representation consisting of both a triangular faceted representation and a complete boundary model. This is the approach taken in the GIVEN (Gesture based Interaction in Virtual Environments) tool-kit (see, e.g., M. Figueiredo, K. Bohm, J. Teixeira, “Advanced Interaction Techniques in Virtual Environments,” Computers and Graphics, v. 17, no. 6, pp. 655-661, 1991) and the ARCADE system (see, e.g., A. Stork, M. Maidhof, “Efficient and Precise Solid Modeling using a 3D Input Device,” Proc. of the Fourth Symposium on Solid Modeling & Applications, Atlanta, Ga., May 14-16, 1997, pp. 181-194). These systems maintain interactive performance by having a “tight coupling” between the input and output handlers, although it is not clear as to how this is achieved. In addition, the authors point out that complex modeling operations necessitate calls to the ACIS™ solid modeler, with a consequent loss of interactivity.
Another possible approach is to simply use a conventional solid modeler for design representation in the VE just as a conventional CAD system does in the 2D environment). Unfortunately, while a solid modeler application can provide most of the required computational functionality for operations in a VR-CAD system (such as modeling, intersection detection, etc.), it is typically cannot be used in such applications because its slowness makes it unsuitable for the real-time requirements of VE-based systems. This conclusion is borne out by the well known O(n2) of basic solid modeler operations such as creation, editing and interrogation of shapes (where n is the number of topological entities, i.e., vertices, edges or faces in a given solid). Another factor influencing the performance of solid modelers is that they typically store a great deal of information in order to maintain consistency of the model (see, e.g., M. Mantyla, “An Introduction to Solid Modeling,” Computer Science Press), which leads to a further reduction in performance.
As noted earlier, a typical solution to modeling issues in VE-based systems is to use a low-level representation of the model coupled to a more detailed representation of its geometry. As noted earlier, the GIVEN and ARCADE systems use a low-level triangular faceted representation in addition to a high-level complete boundary representation. However, as indicated by the authors of the ARCADE system, not all geometric modeling operations can be handled by the high-level representation and therefore a solid modeler is required—but this then results in the possible loss of interactive performance. Therefore, there is still a significant need for methods and apparata which address interactivity problems arising due to modeling issues, as well as the other problems noted above, if efficient and effective CAD systems are to be developed (in particular, effective VR-CAD systems).
The invention, which is defined by the claims set forth at the end of this document, relates to methods and apparata for high-speed computer-aided design of geometric models. As noted earlier, greater design speed (i.e., speed in updating and displaying of the design) generally results in greater interactivity for the user, an advantage which is particularly important if designing is to be performed in computationally burdensome virtual environments (VE). The invention will now be summarized by providing an overview of a preferred embodiment of the invention described in the Detailed Description portion of this document, the preferred embodiment being known as the Virtual Design Software Framework (VDSF).
In VDSF, increased speed is obtained by use of a dual representation of the user-created design, with each representation containing certain information which is useful for modeling the design. These representations may be referred to as the Design Intent Graph (referred to as D) and the Shape Modeling Graph (referred to as S), the term “graph” being used because the representations are best understood when provided in the form of a tree or Directed Acyclic Graph (as will be seen in the Detailed Description section elsewhere in this document). The representations are specially designed to decompose modeling tasks into activities that must be done in real-time and those that can be done asynchronously. Asynchronous performance of modeling tasks may be done in several ways. As a first example, certain modeling data can be pre-computed when features within the design are initially added or edited, and this pre-computed data can reduce the need for calculations later (as will be discussed later, for example, in relation to Geometry Links). As a second example, modeling tasks can be ordered on a spectrum of importance or probability, and those tasks which are more important/likely can be performed first, with less important/likely tasks being performed only if there is sufficient time. This is illustrated later, for example, in the discussion of Proximity Category 1 intersection checks (which are addressed immediately) versus Proximity Category 2 and 3 intersection checks (which are addressed afterwards during processor idle time, if performed at all). Further, in preferred versions of the invention, the representations are constructed so that when action is taken on certain portions of the design, modeling tasks related to that portion are performed locally, rather than over the entire design globally, so as to further decrease computational burden and increase modeling speed. Additionally, the representations are designed to minimize their dependence on conventional solid modeler calculations so that real-time response can be obtained even in computationally-burdensome virtual environments (VE), a result which is not possible if a conventional solid modeler is used.
In the dual representation, the Design Intent Graph D and the Shape Modeling Graph S contain somewhat different information. The focus of D is on storing the “design history,” the sequence of design steps exercised by the user in creating the design. The focus of S is instead on storage of the features of the design with such detail that the design (and the results of editing the design) can be adequately modeled. However, D and S are “linked” insofar as an update of the design representation in D will generally result in a corresponding update in S.
D stores information such as (1) information about the features/elements the user chose for building the design (their shapes, whether they are solid/positive or void/negative elements, the location of element attachments, orthogonal bounding boxes, etc.); (2) the parent-child hierarchy of the elements within the design (a child element being one which is in immediate attachment with one or more prior parent elements); (3) any user-specified or system-specified design constraints on the elements or their relationships (e.g., two elements are to be spaced apart by some specified distance, etc.). While D does not contain sufficient information to fully support all modeling operations that might be performed on the design, it can support rudimentary operations such as constrained relocation (where relocation of an element also relocates that element's children, but does not affect the element's parent); alignment (where an element and its children may be relocated along an axis defined on the element's parent until a parent edge is reached); and simple boundary box-based intersection checks (where intersections between elements are determined by calculating distances between orthogonal boundary boxes surrounding elements within the design).
S is a more complete representation of the design geometry, and is sufficiently detailed to support almost all modeling operations save for a few which are sufficiently complex that they require the assistance of a solid modeler (for example, intersection determinations between boundary representations). S contains information regarding the elements the user chose to construct the design and the relations between these elements, and is most readily understood when presented as a Directed Acyclic Graph (as in FIG. 28, discussed in greater detail later). In such a graph, the elements are represented by nodes and relation information is primarily represented by links between the nodes. Node/element information includes information such as the shapes of the elements, including the faces, face edges, vertices, and characteristic parameters of the elements; information that can later be used to expedite modeling operations, such as the convex hulls of the elements (which can be used for rapid collision detection between elements, as discussed later); and pointers indicating the links applicable to the face of each component. The S representation in VDSF then uses several types of links, which may be summarized as follows.
The Attachment Link A is used where one element within the design is attached to another element. An attachment link may include information such as an identification of the mating/attachment faces between the elements, and the location of each element with respect to others (i.e., an attachment datum or transformation identifying how the attachment face of one element is positioned with respect to the attachment face on the other element). This information is primarily specified by a type of attachment link known it as a primary attachment link Ap. If an element is a void/negative element rather than a solid/positive element, a secondary attachment link As is used in conjunction with the primary attachment link Ap to indicate negativity. The attachment link A information, when used in conjunction with the node/element information, is sufficient to fully describe most designs and support their editing and rendering. As an example, if a child element is moved with respect to a parent element, the attachment link(s) between them are simply updated to reflect the new location of the child with respect to the parent. Any children of the child element will maintain their locations relative to the child element (unless their attachment links are also modified), and therefore changes in the design are localized about the child element. As another example, if a child element is resized with respect to a parent element, the child's node information is updated to reflect the new dimensions/configuration of the child, and the attachment link's location transformation between the child and the parent (and possibly between the child and any children of that child) may also be updated to reflect a change in relative location owing to the child's resizing. Again, changes in the design will be localized about the child unless other attachment links away from the child are altered. Thus, by representing attachments between elements by the attachment links, and by reflecting attachment changes by updating only the attachment links related to the element being edited, the representation of the design is only updated locally; it will not require a more extensive or “global” recalculation. This results in a significant decrease in computation time, allowing the design to be updated and rendered very rapidly so that greater interactivity can be attained.
Further links may then be used if even greater gains in interactivity are desired. The Proximity Link P can be used to expedite intersection (element collision) detection during editing activities by precomputing certain quantities useful in intersection detection, thereby reducing later computation time when intersection detection is required. The P links enable rapid intersection detection by storing all distances d between the convex hull of an element in question and the convex hulls of all other elements which are neither parents nor children of that element. These other non-parent/non-child elements are then ranked in order of increasing distance d (i.e., in order of proximity). When the element in question is later edited (e.g., moved), a convex hull intersection check is only performed between the element and some predetermined number of the most proximal non-parent/non-child elements, and the intersection checks are performed in order of rank. If a convex hull intersection is detected between elements, a “full” intersection check between the boundary representations of the elements can be performed with a call to a solid modeler. Thus, the time-consuming intersection check by the solid modeler is only performed in certain instances when prior tests—whose parameters are largely pre-computed in prior time steps—indicate that an intersection is likely. Additional time advantages are realized if the non-parent/non-child elements are further ranked into three categories:
Proximity Category 1: Elements which have the same parent as the element in question and which are attached to the same parent face as the element in question:
Proximity Category 2: Elements which have the same parent as the element in question but which are not attached to the same parent face as the element in question; and:
Proximity Category 3: Elements which do not have the same parent as the element in question.
In general, this categorization further classifies elements in order of their probability of intersection to the element in question, with intersections being most likely for Category 1 elements and least likely for Category 3 elements. By performing the aforementioned intersection detection test on elements within Category 1 first, and then proceeding to succeeding categories if no intersections are detected, further time savings are obtained because intersections are sought where they are most likely to occur.
If an intersection is detected, the designer can be asked to confirm or deny whether the intersection was intended before being allowed to make further design changes. In VDSF, many proximity links are categorized as “soft” because they allow the designer to permit intersection if he/she wishes to do so. It is also possible to provide “hard” proximity links which do not allow intersections, or which do not permit the elements subject to the proximity link to violate some predetermined proximity condition. As examples, hard proximity links can restrict elements to remain at some predefined distance, or require them to remain above or below some predefined distance. Such hard proximity links can be enforced by checking methods similar to, the aforementioned method of intersection checks, e.g., the distances d can be checked for compliance with predefined distance constraints.
Geometry links G may also be provided to attain further benefits in computational speed. Geometry links are used to store precalculated information regarding the features of elements, or regarding relationships between elements, so that this information is readily available for later use in editing operations. As one example, “parallel face” geometry links may exist between opposite parallel faces within an element so that certain later design activities requiring an opposite parallel face determination (e.g., the piercing of the element by a through-hole) are expedited. Similarly, “opposite face” geometry links may be formed between spaced parallel faces on different elements so that later design activities requiring an opposite parallel face determination (e.g., addition of a “bridge” element) are expedited.
Thus, it is seen that the foregoing design representation can allow several important benefits: (1) it can allow localization of the effect of editing operations (i.e., editing can occur on localized design elements rather than on the design globally); (2) proximity of design elements in the “conceptual space” of the representation can be translated to physical proximity of elements in geometric space to enable faster intersection determination between design elements; and (3) the geometry of design elements can be precomputed so as to avoid potentially expensive computations during editing. These benefits each provide significant computational time savings, which translates to higher interactivity in the design system.
The invention will be better understood upon review of the following detailed description in conjunction with the accompanying drawings, which may be summarized as follows:
FIG. 1 illustrates a functional flow diagram for a prior art CAD system.
FIG. 2 illustrates a functional flow diagram for an exemplary VR-CAD (virtual reality CAD) system.
FIG. 3 conceptually illustrates the structure of the VDSF (Virtual Design Software Framework), a preferred embodiment of the present invention.
FIG. 4 illustrates the Virtual Locator, an input tool used in VDSF.
FIGS. 5A-5C illustrate a series of examples showing use of a Virtual Locator to select, add, and resize elements in a design.
FIG. 6 illustrates a design created via a construction approach, wherein the component elements (all positive entities) are attached by an “Add On” operation to produce the overall model.
FIG. 7 illustrates various primitive shape elements available for use in VDSF.
FIG. 8 illustrates a design analogous to that of FIG. 6, but wherein several elements are provided as negative entities rather than positive entities (i.e., the elements are subtracted rather than added).
FIGS. 9A-9C illustrate valid and invalid “Add On” operations in VDSF for assembly of designs.
FIGS. 10A-10C illustrate a 3D “rubber-banding” operation for element creation and sizing.
FIG. 11 illustrates an example of constrained location editing, wherein editing is limited so that one element may be moved with respect to a second element but cannot be moved outside of the boundaries of the attachment face on the second element.
FIG. 12 illustrates an example of viewpoint dependent editing, wherein input specifying that an element be moved to the right or left side of a second element is executed in accordance with the right and left sides of the elements as they are displayed to the user.
FIG. 13 illustrates an example of a violation of an implicit constraint (design rule) provided in VDSF (non-obstruction of holes).
FIG. 14 illustrates an example of a violation of an explicit constraint (design rule) provided in VDSF (distance between elements does not meet user-specified standard).
FIGS. 15A-15C illustrate progressive steps in the creation of a Design Intent Graph D corresponding to the illustrated design.
FIGS. 16A and 16B illustrate a design and Design Intent Graph D analogous to that of FIG. 15C, but wherein several elements are provided as negative entities rather than positive entities (i.e., the elements are subtracted rather than added).
FIG. 17 illustrates a desired design.
FIG. 18 illustrates a first alternative for constructing the desired design of FIG. 17, with a corresponding Design Intent Graph D.
FIG. 19 illustrates a second alternative for constructing the desired design of FIG. 17, with a corresponding Design Intent Graph D.
FIGS. 20A and 20B illustrate addition of a bridge element (an element spanning at least two other elements but not being attached to a face common with these elements), resulting in the Design Intent Graph D being a Directed Acyclic Graph (DAG) rather than a tree.
FIGS. 21A and 21B illustrate addition of a wedge element (an element spanning at least two other elements and being attached to a face common with these elements), resulting in the Design Intent Graph D being a Directed Acyclic Graph (DAG) rather than a tree.
FIGS. 22A and 22B illustrate the user's definition of an explicit constraint (design rule) wherein the distance between two elements is not to fall below a predefined value, and the representation of this constraint as a link on the Design Intent Graph D.
FIG. 23 illustrates a Design Intent Graph D for the illustrated design, along with the node information contained within certain nodes of D.
FIGS. 24A-24D illustrate examples of location operations enabled by the Design Intent Graph D.
FIGS. 25A-25D illustrate examples of alignment operations enabled by the Design Intent Graph D.
FIGS. 26A-26D illustrate examples of boundary box-based intersection checks enabled by the Design Intent Graph D.
FIG. 27 illustrates an exemplary design and its corresponding Design Intent Graph D.
FIG. 28 illustrates a partial Shape Modeling Graph S (only attachment links A being shown) for the design shown in FIG. 27.
FIGS. 29A and 29B illustrate an exemplary design wherein elements are being edited (moved) therein, and the corresponding Design Intent Graph D and Shape Modeling Graph S (the appearance of the graphs being the same for the model both prior to and after editing, though information stored in the graph links may be changed).
FIG. 30 illustrates an exemplary design along with its D and S graphs, showing the use of primary attachment links Ap in S.
FIG. 31 illustrates an exemplary design having a hole included therein (i.e., having an element subtracted/a negative entity added), along with its D and S graphs, showing the use of secondary attachment links As (as well as primary attachment links Ap) in S.
FIG. 32 illustrates an exemplary design having a hole through two entities, along with its D and S graphs, further showing the use of primary attachment links Ap and secondary attachment links As in S.
FIG. 33 illustrates an exemplary design with its Shape Modeling Graph S, showing the use of soft proximity links Ps (for only element r6) in S.
FIGS. 34A and 34B illustrate an exemplary design with its Shape Modeling Graph S, showing the use of soft proximity links Ps (for only element b5) in S.
FIG. 35 illustrates an exemplary design wherein an explicit design constraint applies (distance between two elements is to exceed a predefined value), along with its D and S graphs, showing the constraint reflected in a hard proximity link Ph.
FIGS. 36A and 36B illustrate S graphs resulting from an intra-face element intersection (i.e., both elements intersect on the same parent) before and after splitting is applied (i.e., before and after the elements are decomposed into new disjoint elements).
FIG. 37 illustrates the Design Intent Graph D resulting from the use of splitting in the model of FIGS. 36A and 36B.
FIG. 38 illustrates the Design Intent Graph D where splitting is not applied in the model of FIGS. 36A and 36B.
FIG. 39 illustrates the Design Intent Graph D where grouping is applied in the model of FIGS. 36A and 36B.
FIGS. 40A-40E illustrate conversion of a soft proximity link Ps to a secondary attachment link As in a (partial) Shape Modeling Graph S in the event where element c7 is elongated (Case I) or carried (Case II) until it pierces element b6.
FIG. 41 illustrates an example of an opposite face geometry link Go listing the opposing faces in element b1 (only a partial list being illustrated).
FIG. 42 illustrates an example of parallel face geometry link Gp listing parallel opposing faces between elements b2 and b3.
FIG. 43 illustrates the resulting change in the Shape Modeling Graph S of FIG. 42 when a bridge element d4 is added between elements b1 and b3.
FIGS. 44A and 44B illustrates a design and its Shape Modeling Graph S, illustrated both with and without relevant geometry links Go and Gp, to illustrate the usefulness of the geometry links in the event a through-hole is extended through all of elements r1, r2 and r3.
FIG. 45 illustrates the current architecture of VDSF.
FIG. 46 illustrates a design of a telephone created in VDSF.
FIG. 47 illustrates the Design Intent Graph D and Shape Modeling Graph S for the button pad p5 (and its nine connected buttons) in the telephone design of FIG. 46.
FIG. 48 illustrates the Design Intent Graph D for the telephone design of FIG. 46 (the buttons of pad p5 not being explicitly shown).
FIG. 49 illustrates the Shape Modeling Graph S for the telephone design of FIG. 46 (the buttons of pad p5 not being explicitly shown), with only attachment links A being shown.
FIG. 50 illustrates the Shape Modeling Graph S for the telephone design of FIG. 46 (the buttons of pad p5 not being explicitly shown), with attachment links A and proximity links P being shown.
FIG. 51 illustrates the Shape Modeling Graph S for the telephone design of FIG. 46 (the buttons of pad p5 not being explicitly shown), with attachment links A and geometry links G being shown.
FIGS. 52A-52C illustrate an exemplary design created in VDSF in conjunction with its Design Intent and Shape Modeling Graphs D and S (only attachment links A being shown in the S graph).
FIGS. 53A-53C illustrate another exemplary design created in VDSF in conjunction with its Design Intent and Shape Modeling Graphs D and S (only attachment links A being shown in the S graph).
FIGS. 54A-54C illustrate another exemplary design created in VDSF in conjunction with its Design Intent and Shape Modeling Graphs D and S (only attachment links A being shown in the S graph).
FIGS. 55A-55C illustrate another exemplary design created in VDSF in conjunction with Design Intent and Shape Modeling Graphs D and S for certain of its subcomponents (only attachment links A being shown in the S graph).
FIGS. 56A-56B illustrate construction of another exemplary design created in VDSF.
FIG. 57 illustrates a hypothetical design that a user wishes to create.
FIGS. 58A and 58B illustrate creation of the design of FIG. 40A by use of solely “Add On” operations, whereby surface attachment of elements is used to attain the final model.
FIGS. 59A-59C illustrate creation of the design of FIG. 40A by use of an “Add In” operation provided in alternative embodiments of the invention, whereby direct insertion of one element into another is used to attain the final model.
As previously noted, the invention relates to methods and apparata for rapid design of objects in CAD environments. The invention will now be described in terms of an actual working embodiment which has been developed in the I-CARVE laboratory of the University of Wisconsin (Madison, Wis., USA), and which will be referred to as the Virtual Design Software Framework, or VDSF. While the following discussion of the VDSF will illustrate how to make and use the invention, it must be kept in mind that the VDSF is merely a preferred version of the invention, and the full scope of the invention is defined by the claims set forth at the end of this document.
The VDSF allows a user to design (i.e., create, edit, visualize, and manipulate) objects, including extremely complex objects, very rapidly in a three-dimensional VE by use of motion (hand gesture) and voice input. The VDSF can be viewed in terms of several components or “layers” which are schematically illustrated in FIG. 3:
(1) An Interaction Mechanisms Layer (or User Interaction Layer), which allows the user/designer to interact with the VDSF via input/output devices (e.g., tracking devices and sound transducers for motion and voice input, and monitors and speakers for visual and sound output). A WISL (Workspace-Instances-Speech-Locator) paradigm is used to it allow designer interaction with the VDSF, wherein designs are created in a 3D Workspace by Instancing simple shapes using Spoken commands and interactively Locating them in 3D space using hand input.
(2) A Design Editing Layer which allows the user to act on objects in the Virtual Modeler Layer (discussed below) via the Interaction Mechanisms Layer.
(3) A Virtual Modeler layer, which provides the underlying geometric representation of the design. This representation is provided in two parts:
(a) In a Design Intent Graph (D graph), which stores the faceted primitive elements that are combined to assemble a design, and additionally stores the hierarchy in which these elements were combined (i.e., the sequence of commands used to assemble the elements in the design). The Design Intent Graph D is the least detailed representation of the design contained within VDSF, but it allows high-speed modeling for a large number of designs (those utilizing only “positive” elements) and editing operations (such as relocation of elements, alignment of elements, etc.).
(b) In a Shape Modeling Graph (S graph), which provides a more detailed representation of the design than the Design Intent Graph D. The Shape Modeling Graph S describes a design in terms of the boundary representations of the basic shape elements making up the design (these boundary representations being obtained from the Solid Modeler Layer discussed below), and also in terms of links which contain information regarding the relationships between elements. The S graph provides high-speed modeling in certain cases when the D graph is insufficient, such as when designs include negative elements and when certain editing operations are required that cannot be accommodated in D (automatic dimensioning, etc.).
The unique dual-graph structure and methodology of the Virtual Modeler Layer allows exceptionally rapid design in a VE by allowing a designer to quickly and easily set forth his/her design intent, with certain nonessential modeling, intersection detection, and rendering steps occurring asynchronously so that the time-critical demands of a true VE environment can be met. As will be evident from the discussion presented later in this document, the Virtual Modeler Layer offers a significant advance in providing true interactivity in VE environments (though many of its advantages can be realized in conventional two-dimensional systems as well).
(4) A Solid Modeler Layer, wherein a solid modeler functions as a geometry engine capable of responding to queries from the S representation. The solid modeler supplies the S graph with local boundary representations (those of its elements), and can also be called upon to generate a global boundary representation of the entire design in instances where the S representation may prove insufficient for modeling.
Each of these layers will now be discussed in turn in greater detail.
As mentioned earlier, speech; and hand input are the primary modes of user interaction in VDSF. The WISL paradigm allows the designer to become part of a 3D VE workspace in which he/she creates instances of primitive shapes using spoken commands and specifies positions of these shapes using a motion controlled 3D locator. The VE workspace therefore replaces the “windows metaphor” commonly found in CAD systems, with speech input replacing “icons” and “menus”, and the 3D locator playing a role analogous to a 2D mouse in a typical CAD system. This paradigm emulates the real world, in which a designer will often use a combination of spoken words and hand motion to indicate approximate sizes and shapes of products during the conceptual design stage.
As opposed to several prior computer systems utilizing motion input, gestures having specialized, preassigned meanings are not used in VDSF since it was felt that such gestures are not necessarily the best way to communicate design intent. This is because using gestures would require training, which could hinder VDSF's use. Instead of gestures, hand motion input is used in intuitive fashion as a 3D locator that is used for positioning shapes created via voice commands. While several modes of hand motion input are possible (and are discussed in greater detail in other references, e.g., C. C. Chu, T. H. Dani and R. Gadh, “Multimodal interface for a Virtual Reality Based CAD System”, CAD Journal, March, 1997), the preferred mode of hand motion input in VDSF is the Virtual Locator shape manipulation tool, an example of which is illustrated in FIG. 4. The Virtual Locator is a hand-held pointer with one or more buttons, and its real-world orientation is represented on a display as a line extending from a virtual hand (as illustrated in FIGS. 4A-4C, which are discussed at greater length below). By pointing the Virtual Locator and issuing commands via its button(s), voice input, or other forms of input, the Virtual Locator can be used for navigation or for object selection, editing and manipulation. As an example, the Virtual Locator allows navigation in the VE (i.e., it allows going to desired areas in the virtual workspace) by pointing to an object or area and then issuing a spoken “zoom-in” command. FIGS. 5A-5C then illustrate use of the Virtual Locator for editing purposes. FIG. 5A shows the Virtual Locator indicating a target face on an object on which some operation is to be performed. Once the face is chosen, the Virtual Locator is used to position other objects/primitives such as the cylinder, shown in FIG. 5B. The Virtual Locator can also be used to interactively modify the geometry of a selected primitive. FIG. 5C shows the locator being used to rescale the size of the previously added cylinder.
The Design Editing Layer has two primary components: (1) an Interaction Component that allows navigation, selection and manipulation of the VE environment (which are subsidiary but useful tasks in the VE); and (2) a Modeling Component which allows modification of object geometry. Each of these components will now be described in greater detail.
(1) Interaction Component: The Interaction Component provides different methods of navigation in 3D space (getting to various locations in the VE workspace); selection of objects (for example, by pointing); and manipulation of those objects (for example, by grasping an object by hand and interactively rotating it). Since most designers are accustomed to 2D windows in a typical CAD environment, navigation can be difficult in a 3D VR-CAD system unless it is highly interactive. In VDSF, the Virtual Locator concept is used in conjunction with motion input and spoken commands to allow the designer to point to an object and then issue a spoken command, e.g., “zoom in” to get the desired zoomed-in view.
Object selection via motion input requires real-time 3D-intersection detection between the Virtual Locator and the objects in the VE. Once an object is selected, it can be manipulated by the designer to get a particular view. Here, manipulation refers to activities such as picking and placing an object, or grasping and rotating an object to get a desired view. It also requires appropriate geometric transformations so that the user is able to pick and place objects in a natural fashion. Visual output generation requires tessellation of geometric models into triangles or other forms needed by graphics libraries, and computation of a pair of images with parallax separation to enable a stereoscopic visual display. Additional geometric computations are necessary to provide auditory and/or haptic feedback. For example, haptic feedback (if provided) requires computation of surface properties so that the appropriate texture can be simulated in the haptic feedback devices.
(2) Modeling Component: VDSF's WISL interaction paradigm bases its design modeling tools upon use of speech commands and hand input. An alternate approach could have instead adapted existing geometric modeling tools in conventional WIMP (Windows-Icon-Menu-Pointer)-based CAD systems to the WISL paradigm. For example, modeling tools based on sketching extrusions and then adding geometric features to the resultant geometry could be provided with hand input (for use in sketching) and speech input (for used in adding instances of geometric features at specific locations). However, not all conventional modeling tools can be readily adapted for use in a VE using a WISL paradigm. For example, Constructive Solid Geometry (CSG)-based approaches wherein primitives are combined via union, subtraction and negation operations may be difficult for the user to specify via WISL as opposed to textual approaches (as in some CAD applications). Thus, modeling tools in existing CAD systems may not be appropriate for use with WISL and similar methods.
The current approach in the VDSF is based on a “construction” approach similar to CSG operations, which allow creation of complex shapes by aggregating prede-fined shape elements. However, VDSF's WISL paradigm leverages the benefits of speech and motion input interfaces to simplify editing, and does not rely on a direct translation of WIMP operations to WISL operations. VDSF's approach is best understood with initial reference to an example. Consider the compound block model in FIG. 6. To create this S model, the voice commands (indicated in quotes) and the hand input are as follows:
(1) Create an instance of a block (b1) with default parameter values. In general, the designer inputs the default parameters using spoken commands; this approach is discussed further, e.g., in T. H. Dani, R. Gadh, “Creation of Concept Shape Designs via a Virtual Reality Interface,” CAD Journal, Elsevier Publications, 29(7), 1997.
(2) Point to the appropriate location on the top face of b1 using the Virtual Locator and “Add On Boss” (o2) by voice input.
(3) Point to the appropriate location on the top face of o2 and “Add On Boss” (o3) by voice input.
(4) Point to the appropriate location on the top face of b1 and “Add On Rib” (r4) by voice input.
Thus, simpler shape elements (some of which are shown in FIG. 7) are combined to create more complex shapes. The Virtual Locator is used to select a desired face of an element and to locate a further element to be added. The elements are typically created by a speech command. This approach has the advantage that it does not require 3D positioning of shape elements before “Adding On” one element to another, an approach which could be tedious. The representation in the Virtual Modeler Layer (to be discussed in greater detail elsewhere in this document) allows such a construction and supports the requirements of the WISL interface.
The primitive shape elements in FIG. 7 reflect the commonly occurring geometric features in mechanical engineering components (see, e.g., P. R. Wilson, M. J. Pratt, “A Taxonomy of Features for Solid Modeling,” Geometric Modeling for CAD Applications, M. J. Wozny, H. W. McLaughlin, J. L. Encarnacao (Eds.), Elsevier Science Publishers, 1988, pp. 125-137). Each shape element (si, where i represents the ith shape element in an object) within an object in VDSF denotes the presence or absence of material. For example, rectangular blocks (with appropriate parameter values), when created as negative entities, may be used to “Add On” rectangular slots, pockets and holes to an object (as shown in FIG. 8, wherein the “basic” object to which the slots, pockets and holes are added is itself a “positive” rectangular primitive). The library of shape elements shown in FIG. 7 can be extended as desired to include other possible primitives having different shapes. Each primitive is modeled as a collection of surfaces with vertex, edge and face adjacency.
“Adding On” si to sj requires that si and sj share a pair of mating faces (which are either partially or fully overlapping). As a consequence, geometric configurations of the type shown in FIG. 9A and FIG. 9B are acceptable, but the configuration in FIG. 9C where an edge of block b2 touches the face of block b1 is disallowed since it results in the creation of a non-manifold model.
Other modeling operations in VDSF allow modification of preexisting geometry, and these include transformations such as translations, rotations and scaling. In addition, 3D rubber-banding (which behaves like the 3D equivalent of 2D rubber-banding techniques allowing interactive sizing of objects) allows interactive creation of shapes via hand input. This includes (1) creation of free-standing shape elements, i.e., shapes not attached to any other shape elements, and (2) in-place addition of shape elements to an existing element (as shown in FIGS. 10A-10C). In both cases, hand input is used to interactively size the element until the desired size is reached. This approach allows rapid creation of shape elements by specification of approximate dimensions via hand input. Alternatively, as previously noted, exact dimensions can be specified via the voice-input mechanism.
Given that exact location and editing of shapes in three dimensions is difficult, the Design Editing Layer provides various types of constraints (design rules) that allow simplification of interactive placement and shape modification. Design constraints may be grouped into two categories:
(1) Implicit constraints, i.e., those design rules that are automatically generated by the system or are default constraints. One exemplary such constraint in VDSF is that once si is added to sj, si is constrained to lie on sj even as it is interactively edited. Such editing is shown in FIG. 11, where r2 is added to b1 and afterwards may be moved in either direction. During this re-location, r2 continues to lie on the surface of b1 in accordance with the aforementioned constraint. Another type of implicit constraint is a viewpoint-dependent location constraint which allows rapid alignment of si with respect to sj (i≠j). An example of such a viewpoint dependent location constraint is illustrated in FIG. 12, where a designer issues a command for r1 to move to the “Left Edge,” and the system responds by relocating the rib to the left edge taking into account the viewpoint of the designer. Another implicit constraint, non-obstruction of predefined negative elements, is illustrated in FIG. 13, where the designer is not allowed to move rib r5 to obstruct hole (negative element) h4. Automatic geometric constraints such as this one could be relaxed, but are preferably provided to allow simplified editing operations.
(2) Explicit constraints, i.e., those design rules that are explicitly specified by the designer. FIG. 14 illustrates an explicit constraint violation: “Ribs to maintain a minimum distance of d*”, where the distance d* between the two ribs created by the designer becomes less than that allowable by a constraint specified by the designer owing to design/manufacturing/assembly requirements. Unlike implicit constraints, which are built into the Design Editing Layer, explicit constraints can be redefined by the designer.
The Virtual Modeler Layer provides VDSF's underlying geometric representation of objects, the data within it is organized in such a manner that it may be represented as a dual graph consisting of a Design Intent Graph (D) and a Shape Modeling Graph (S). The dual graph representation assists in accomplishing two goals:
(1) To capture the designer's design intent, including collection of the designer's modeling steps used to create a design and recording the specified design constraints so that this information can be used in future design activities. This is primarily done within D. It must be noted that the notion of “design intent” in D refers to the intended/desired geometric relations between the model's features rather than the actual functional intent of the design, i.e., D is related to capturing the history of how the shape was created rather than to the design theory and end operation of the model.
(2) To provide a geometric representation for modeling and graphical rendering of created models within the time-critical requirements of the CAD system (in particular, in VR-CAD systems wherein design occurs within a VE); in other words, to allow modeling to occur with “true interactivity.” This is done by using the less detailed D representation to perform certain basic modeling chores, and then relying on the more detailed S representation when the complexity of the modeling step at hand is such that the D representation cannot adequately perform it. However, even the S representation—which relies on the solid modeler for a portion of its operation—is optimized in such a fashion that it achieves much faster performance than a solid modeler.
Note that while D and S are distinct, they are related insofar as each shape element in D is “mapped” or “coupled” to one or more entities in S. The two separate design representations of D and S are used because D is not well suited for certain types of design steps and their graphical display. While the structure of D makes it possible to use it for the display of relatively simple design geometries (such as designs composed of positive shape elements combined with “Add On” operations), it cannot be used for display of shapes containing negative entities, and additionally it cannot fully accommodate more advanced modeling functions (e.g., resizing of shape element geometries). In these cases, S must be used since it is capable of representing shapes containing both positive and negative entities, and it can handle more advanced modeling operations because it contains a more detailed representation of the design. S uses a hierarchical graph structure that represents a design in terms of disjoint basic shape elements (the geometry) connected via various links (the topology). These links include attachment, proximity and geometry links which not only store the relationships between the elements but also enable faster computations in S (as will be explained later), thereby providing interactivity advantages. Each of the Design Intent Graph (D) and Shape Modeling Graph (S) will now be discussed in greater detail in turn.
The Design Intent Graph D directly reflects the sequence in which the designer creates the design of the geometric model, and captures the design history based on the steps taken by the designer to create the shape. The D graph provides a high-level representation of the design which allows many common modeling functions (i.e., common shape element creation and editing functions) to be performed without requiring queries to the solid modeler. As a result, modeling functions that can be accomplished with sole reference to D can be executed extremely rapidly.
During the design process in VDSF, the designer specifies the geometry of the design in terms of the shape elements described earlier. For example, if the designer wishes to specify a design feature such as a rib, this is geometrically equivalent to a positive rectangular block element with a width considerably less than its length. Similarly, a slot is a negative instance of a rectangular block element with appropriate dimensions. The designer's specification of various shape elements Fi results in D—a directed graph—wherein the elements Fi appear as nodes of the graph. A directed link on the graph D from element Fj to Fi (Fj→Fi) indicates that the designer has added Fi to Fj (thus Fi is a “child” of “parent” Fj). To illustrate by example, FIGS. 15A-15C show the successive steps in the creation of an example design and the resulting Design Intent Graph D at each step. FIG. 15A illustrates the addition of block b2 on b1; followed by boss o3 on b2 (FIG. 15B); and finally rib r4 on b1 (FIG. 15C). FIG. 16A then illustrates a design containing the “negative” counterparts of the elements b1, b2, o3, r4 of FIG. 15C, added to a rectangular block element in the same order as in FIG. 15C. Its corresponding D is shown in FIG. 16B. Note that while D for the design in FIG. 15C is structurally similar to that of FIG. 16B, the geometry of the two designs is significantly different.
It is seen that the resulting Design Intent Graph D is a data structure commonly referred to as a tree or Directed Acyclic Graph (DAG). As a result, D will generally have the following properties:
(1) For a large variety of cases, such as the ones in FIGS. 15 and 16, D assumes a simple tree structure. Advantageously, typical search algorithms within a tree structure have complexity O(log n), where n is the number of nodes in the tree. For some special cases, D might have a non-tree structure containing one or more cycles (i.e., a true DAG), as will be discussed later.
(2) Two designs with the same geometry may have distinct D's depending upon how they are constructed, as illustrated in FIGS. 18 and 19. The desired design of FIG. 17 may be constructed in the fashion of FIG. 18, wherein a slot (a negative block element of suitable dimensions) is added to an existing larger block. Alternatively, it may be constructed as in FIG. 19, wherein two blocks of a smaller size are added to an existing block to create the same shape. The different design steps for each case are reflected by the D's shown in the respective figures. While the two final shapes are identical, the intent of the designer in each case is different, and therefore D is different.
(3) Similarly, two designs having different geometry can have similar D's. An example was noted above regarding FIG. 16, wherein several elements in the design are the negative versions of the elements of FIG. 15. The D for the design of FIG. 16 was seen to be very similar to that of FIG. 15.
These examples illustrate that the construction approach for constructing D, and the design history implicit in the construction, will have an impact on future methods of editing designs. For example, if the model in FIG. 17 is constructed as shown in FIG. 18, future editing operations will involve changing the dimensions of the slot b2. For FIG. 19, editing the width of the “slot” created by the two blocks b1 and b2 will require modifying the dimensions of either b1 and/or b2.
D for the above examples is a tree structure. This is because each Fi is added to a single existing Fj, and therefore each Fi has a single parent (which is necessarily true for a tree). While this construction is useful for most designs, in some cases it may be necessary to add a child element Fi to more than one parent element (Fj and Fk) simultaneously. Examples include designs with elements that serve as bridge elements between two other elements, as shown in FIGS. 20A and 20B, or wedge-type elements that can be added to pre-existing geometry, as shown in FIGS. 21A and 21B. For the former, FIGS. 20A and 20B show block b4 added to both blocks b2 and b3. For the latter, FIGS. 21A and 21B show block b3 added to both b1 and b2. In such cases, D is no longer a tree but a true DAG. In a DAG, it is not possible to make a cycle in the graph when following the directional links; further, it is possible to have more than one path from node F1 to Fj (i≠j). While the former characteristic is true for a tree, the latter is not, and hence a DAG is more efficient than a generalized graph, though less efficient than a tree (see, e.g., A. V. Aho, J. E. Hopcroft and J. D. Ullman, Data Structures and Algorithms, Addison Wesley Publishing Company, 1983).
Another role of D is to store the design rules/constraints specified by the designer while creating the design for which the rules apply (examples of such constraints being noted earlier in the discussion of the Design Editing Layer). When a constraint is specified, an instance of it is created and stored as a part of D in the form of an explicit link between the set of shape elements to which it applies. For example, in FIG. 22A, the designer specifies a minimum distance x* between any two ribs ri and rj (i≠j). This constraint is stored in the form of a special link (shown dotted in FIG. 22B for only ribs r3 and r4) and becomes part of the Design Intent Graph information. As with the other portions of D, this link is translated into an appropriate link in S for use in rendering the graphical display. It has been observed that most common design rules/constraints are based on only two shape elements, and VDSF applies this as an assumption. However, VDSF could be modified to allow constraints which extend to three or more shape elements.
While the links in D capture the parent/child hierarchy of shape elements and any design constraints concerning the elements, the nodes contain additional geometric information, as illustrated in FIG. 23. The node information includes the name of the element, the type of element, a transformation locating the element with respect to the parent(s), the parameters for the element, a triangular faceted representation of the shape element, and a bounding box aligned with the parent's coordinate system and surrounding the element. While the name, type, and location of the element are generated from user input, the faceted representation is obtained via S (which itself receives the representation from the shape modeler). Once computed, the faceted representation does not require updates unless the dimensions of the element are changed during subsequent interactive editing.
D is capable of fully representing designs in which simple editing has occurred; in other words, the D domain only reflects editing operations which require minimal updates to the detailed geometry of the shape element stored in S. As will be discussed below, simple editing commands fully supported by D include positioning commands such as constrained location, alignment and bounding box-based intersection checks during geometry editing. While the constrained location and alignment commands provide a quick way to position shape elements, the bounding box-based intersection checks provide the ability to detect potential collisions between elements.
The Constrained Location operation is illustrated in FIGS. 24A-24D. Here the constraints are the constraints on the transformations imposed by the parent-child relationship between two elements. For example, in FIG. 24A, b2 is interactively moved to a new location and child b3 stays attached to it. Conversely, the motion of child b3 is independent of its parent b2, as shown in FIG. 24B. FIG. 24C shows that the hierarchy is preserved even if b2 is moved off the face of b1 to which it was originally added. The ability to relocate b3 independently of b2 still remains, as shown in FIG. 24D. In each of these examples, the transformation applied to an element is propagated to all its children, but not to its parent. The constrained motion operation can be achieved by requiring that a selected element follow the intersection point of the Virtual Locator with the faces of the design.
The Alignment Operation provides another means to rapidly relocate elements apart from manually relocating them, with FIGS. 25A-25D illustrating element positioning via alignment operations. FIG. 25A illustrates alignment of b2, a child of b1, in a +X axis fixed on b1. The child b3 of b2 is carried with b2. Similarly, FIG. 25B illustrates alignment of b3 in the +X axis fixed on its parent b2. FIG. 25C illustrates alignment of b3 with its parent b2 in the −Y axis direction, and FIG. 25C illustrates alignment of b2 with its parent b2 in the −Y axis direction. Note that since it is generally difficult for a user to quickly ascertain the coordinate system on an element (e.g., which direction on an element is defined as the +X direction, which is defined as the −X direction, etc.), VDSF uses a coordinate-system independent, user-viewpoint-dependent method of alignment whereby the user issues intuitive commands such as “Align with left edge of [parent element],” “Align with right edge of [parent element],” etc. This utilizes the viewer's current viewing orientation and the orientation of the specified parent's coordinate system to determine the directions corresponding to right, left, front, rear, top, bottom, etc. In other words, the VDSF display viewed by the user is considered as having a right-left/top-bottom/front-rear coordinate system, and when a viewpoint-dependent alignment command is issued (e.g., “Align with right edge of [parent element]”), VDSF determines which of the XYZ coordinate axes fixed on the parent most closely corresponds to the right-left axis of the display. The child element is then moved on the parent along that axis of the parent.
The Intersection Operation performs computationally inexpensive bounding box-based intersection checks between siblings of a single parent. As previously noted, when an element is added to D (when a node is formed in D), a bounding orthogonal box is defined about the element and stored in the node. The distances between spaced bounding boxes of different elements are quickly and easily calculated because all walls are orthogonally oriented. FIG. 26A and 26B illustrate the case where a positive element is rotated, causing it to interfere with another element; this intersection is readily detected with a bounding box-based intersection, check. Since both positive and negative entities are represented in essentially the same manner in D, bounding box-based intersection checks can also be used to detect whether the slot s5 of FIG. 26C, once rotated, is obstructed by element b2 or r4 in FIG. 26D. This would be indicated if the distances between two or more orthogonal planes on the elements' bounding boxes were zero or less: for example, the orthogonal distance between the top face of s5 is zero with respect to the bottom face of r4, and if the orthogonal distance of another face of s5 was zero or less with respect to another face of r4—say, if the orthogonal distance of the left face of s5 was zero or less with respect to the right face of r4—slot s5 would be obstructed. Naturally, a bounding box intersection check is not always exact since bounding boxes may intersect where actual element boundaries do not, but it generally has the benefit of being “conservative” insofar as it will report false intersections more often than false non-intersections. As will be discussed later, more detailed intersection checks can be performed by the S representation; if desired, these can be triggered by a boundary box intersection in D.
An analysis of the operations possible in the D representation show that excluding operations wherein D needs to call to S and the solid modeler (such as for the faceting of the elements in D), all operations are essentially constant time operations (see TABLE 1). Intersection detection could be a potentially expensive operation if utilized continuously during interactive editing.
SUMMARY OF COMPLEXITY OF OPERATIONS IN D
(N-number of elements, M-number of faces/element)
Operations in D
Locate intersection point
Bounding box intersection check
As noted above, D, typically a tree/DAG structure containing symbolically represented shape elements, has the principal purpose of representing the sequence of design intent steps. While it may also be used for display of relatively simple design steps, its lack of detailed topological relationships between elements and its inability to adequately represent negative entities limits its use for display of many complex design steps of practical interest. Thus, the VDSF also uses the Shape Modeling Graph S, which is significantly more detailed than D and which supports the geometric representation and graphical display of more complex design steps. While being more detailed that D, S is less detailed than the mathematically complete representation provided by a solid modeler, and is more suitable for high-speed interactive editing than the representation contained within a typical solid modeler. To accomplish this, S preserves the parent-child hierarchy of shape elements in D by utilizing nodes representing shape elements, but it provides solid modeler capabilities by storing a boundary representation for each shape element at each node. Maintaining a boundary representation of each element at the local (element) level, rather than at a global (design) level, results in localization of the operations for which S must call to the solid modeler, with accelerated performance in S as a result. S additionally stores (precomputes) auxiliary information between pairs of shape elements si and sj (i≠j) for later use to further speed up interactive design. When shape elements in D are manipulated, explicit pointers map from each Fi in D to the corresponding si in S, and the graphics display is updated appropriately.
As its name implies, the Shape Modeling Graph S models a shape as a graph in which the nodes si (i=1 . . . N) represent the shape elements of the design and the links Ij,k (j=1 . . . N, k=1 . . . N) represent the topological relationships between them. The links/topological relationships between a pair of elements si and sj (i≠j) are of three types: attachment (A), proximity (P), and geometry (G). Of these, the attachment link Aj,k (j=1 . . . N, k=1 . . . N, j≠k) between sj and sk typically arises from the “Add On” operation specified by the designer when adding sj to sk (or vice-versa). While the attachment link A is typically based on the designer's input (though examples to the contrary will be given later), the proximity link P and geometry link G are “derived” links that are not based on D, but are instead based on the specific geometric characteristics of the design being created. Proximity links Pj,k (j=1 . . . N, k=1 . . . N, j≠k) join pairs of shape elements sj and sk and are used to detect intersections between sj and sk during interactive editing. Geometry links Gj,k (j=1 . . . N, k=1 . . . N, j=k is valid) contain pre-computed geometric information that enables design modeling without requiring extensive computations during interactive editing.
An important goal of the dual graph representation in the Virtual Modeler Layer is to minimize the complexity and number of computations that are needed during interactive editing so that a higher degree of interactivity can be attained. The lesser the number of geometric computations that need to be performed during any given time-step, the greater the degree of interactivity. The Virtual Modeler Layer reduces the computations required by pre-computing geometric information that is anticipated to be required during interactive editing. Also, the geometric information available in a given time-step is used to simplify computations in the next time-step, the assumption being that there is an incremental change in the model from one time-step to the next. The Virtual Modeler achieves these functions by use of the A (Attachment), P Proximity) and G (Geometry) links, which are explained in the following sections.
The geometric information in S is composed of boundary representations of the individual nodes si, i.e., the various entities making up the overall design. The topological structure of S preserves the hierarchy of shape elements in D and includes additional information which takes the form of Attachment A, Proximity P, and Geometry G links between nodes si and sj. To initially illustrate the nodes and links in S, the D and S graphs for an exemplary design are shown in FIGS. 27 and 28. As shown in FIG. 27, the design includes two stacked positive block elements b1 and b2, and a negative cylinder element h3 forming a hole in block b1. The Design Intent Graph D is also shown for this arrangement. FIG. 28 then shows the corresponding Shape Modeling Graph S. In reality, only a partial graph S is shown; for clarity, some element/node information and links are omitted, as will be evident from the following discussion. Within S, each element/node si (block elements b1 and b2, and cylinder element h3) is represented by a list of the faces fk (k=1 . . . F), edges ek (k=1 . . . E), and vertices vk (k=1 . . . V) that bound the element, their topological interrelationships, and a list of parameters appropriate for that type of element (exemplary parameters being illustrated in FIG. 7 in conjunction with the elements they characterize). (Note that only the node structure for elements b1 and h3 is illustrated for sake of clarity.) Each face fk (of each element si) is cross-referenced to its relevant A, P, and G links Ii,j (and vice-versa) by pointers. Thus, the positive block b1 is represented by its boundary representation, parameters, and pointers to links that originate or terminate at it (only certain attachment links Ap 1,2, Ap 1,3 and As 1,3 are shown; P and G links, as well as the meaning of the attachment link terminology Ap and As, will be discussed later). The hole is formed in block b1 by adding the negative cylinder h3, which has a top face f1, a bottom face f2, and a cylindrical side face f3. While f2 and f3 form part of the hole representation, they are not displayed in the design since they are part of a negative entity; however, f1 is displayed since this face of the hole is visible. The fact that h3 is a negative entity is stored in the Ap 1,3 link and this information is used when the top face of b1 is displayed (i.e. it is used to denote that f1 of b1 has a circular hole in it). Similarly, the As 1,3 link enables the correct display of the bottom face f6 of b1.
The hierarchy in S, which is derived from D, enables interactive positioning of shape elements in space as well as localization of editing operations (i.e., computations can be performed on specific portions of the design, rather than on the entire design globally). This hierarchy simplifies the task of interactive repositioning of elements, particularly when coupled with constraints provided in the Design Editing Layer. Of these, two constraints are particularly helpful:
Constraint 1: Requiring that a child node stays attached to the parent node, i.e., changing the position of a parent node causes all child nodes to move along with the parent. (However, the reverse is not true, i.e., child elements can be positioned independently of parent elements.)
Constraint 2: The motion of a child element is restricted such that the child element lies on the face of the parent element. This face is typically the one to which the child element is attached.
These constraints are illustrated in FIGS. 29A and 29B. In FIG. 29A, block b2 (a child of b1) is interactively repositioned (as shown by the arrow in FIG. 29B). Block b2 is constrained to stay in contact with b1 at all times, and its child (blind hole h3) moves with it. However, its parent b1 is not affected.
The attachment link A is the fundamental link in S that is used to represent the addition of an element si to an existing element sj in S. The function of A in S is to enable geometric representation of the design, as well as its graphical rendering, without explicitly performing extensive boundary evaluations at every time step (as is done in conventional solid modelers). The attachment links therefore assist in enabling rapid graphical updates and interactive editing of designs.
An attachment link A between elements/nodes si and sj contains the following information about si and sj:
(1) The specific mating/attachment faces of si and sj (fsi and fsj respectively, fsi being the face of si on which sj was added, fsj being the corresponding mating face of sj);
(2) A transformation that locates si on face fsj, used to position si on fsj.
(3) The type of A link, a primary attachment link Ap or a secondary attachment link As.
The mating faces fsi and fsj are implicitly defined by the addition of si to sj, since this is typically specified by the designer when adding si to sj. The transformation locating si on face fsj is based on the position of si indicated by the designer's input. The type of link between si and sj in S is determined by the existence of a corresponding directed link between Fi and Fj in D (assuming that s, maps to Fi and sj maps to Fj). Typically, when one feature Fi is added to another feature Fj, there exists only one such directed link between Fi and Fj in D which forms the basis of the primary attachment link Ap between si and sj in S. This link Ap in S is called a primary attachment link since this is the “primary” link by which si is attached to sj, as specified by the designer (and as reflected in D). Note that for designs containing certain types of elements, such as bridge and wedge type elements (examples of which were given earlier), more than one Ap exists.
Some examples of primary attachment links Ap for a sample design are shown in FIG. 30. In this example, the designer adds b2 to b1, and this is reflected in D. The link Ap 1,2 between b1 and b2 in S captures this relationship between the mating faces of b1 and b2, and is a primary link since the addition of b2 on b1 is specified explicitly by the designer. A similar link Ap 1,3 for b3 on bi is also shown. Based on these examples, it is seen that in general there exists one Ap i,j per pair of mating faces of two shape elements si and sj.
As opposed to the primary attachment link Ap, which is directly derived from D, the secondary attachment link As is system-generated by the Virtual Modeler Layer. Such system-generated links are needed to accurately model the piercing of positive shape elements by negative entities such as holes. These links are called “secondary” since they are typically not specified by the designer (and hence have no analog in D), and are derived based on the geometric properties of the elements in S. As must be computed in order to display certain types of negative entities, such as the through-hole h3 in b1 shown in FIG. 31. The advantage of As is essentially the same as Ap, i.e., it allows local computation (it avoids recomputing the boundary of the entire model during each time-step), while ensuring the accurate graphical rendering of models with negative entities.
In FIG. 31, primary attachment link Ap 1,3 reflects the intent in D (hole h3 added to b1). As 1,3, a secondary attachment link, is a system-generated to capture the piercing of the through-hole on the bottom face of b1. The presence of As 1,3 allows the bottom face of b1 to be correctly graphically rendered (i.e., with a rectangular hole). This is achieved by local computation, i.e., without having to globally recompute the boundary of the entire solid model of the design. (Other links which reduce computational burden and which are relevant to the modeling of through-holes, such as the geometry links G, are explained later.)
While the previous example illustrated the construction of the secondary link for a through-hole passing through one element, the example of FIG. 32 illustrates the addition of a through-hole h4 that pierces two elements b1 and b2. The piercing of the two faces of b1 (the top and bottom faces) is represented by the two secondary links between h4 and b1 (As 1,4 and As 1,4 respectively). Since the hole h4 visibly pierces the top face of b2, primary link Ap 2,4 is included, and since h4 also pierces the bottom face of b2, an additional secondary link As 2,4 is formed in S.
Proximity links are primarily used to enable the detection of intersections between si and other element(s) sj (i≠j, i=1 . . . N, j=1 . . . N) during interactive editing of si. In the VDSF, intersections between elements, once detected, may be allowed or disallowed by the designer. The rationale for doing this is that when a geometric element (e.g., a rib) is added to a prior element (e.g., a block), it is added to serve a specific function. More often than not, the designer will not want the rib to intersect any other feature on the block; the designer may want it to attach to another element (have the surfaces of the elements mate), but he/she will generally not want one to “pierce” the other. This implies that if si and sj are created, and if interactive relocation of si or sj results in a potential intersection between them, such intersections should be detected. Once an intersection is detected in VDSF, the designer may choose to allow or disallow the intersection, and D and S are appropriately updated.
Ideally, such intersections should be detected in real-time so that graphical computation of the edited geometry and visual feedback to the designer can be (practically) instantaneously provided, thereby advancing the goal of increased interactivity. In S, this high-speed intersection detection is done via proximity links Pi,j, which are formed between all pairs of elements si and sj wherein i≠j, si is not a parent of sj, and sj is not a parent of si (i.e., between all element pairs wherein the elements are not in a parent-child relationship). Pi,j stores the minimum distance di,j between the convex hulls hi and hj of si and sj for the purpose of high-speed distance checking. As will be discussed later, determining the minimum distance between the convex hulls of objects involves significantly less computational burden than performing an actual intersection determination between the objects themselves. The preferred method of intersection detection is as follows:
(1) Convex hull distance calculation: At any given time-step, all di,j's (the minimum distances between convex hulls) are computed between the element si currently being edited and all other elements sj which are neither parents nor children of si. The sj's (i≠i) are then ranked based on an increasing value of di,j. This ranking is stored in si (i≠j) so that if si is further edited at the next time-step, only the most proximal sj candidates from this ranking (those with lowest di,j) are checked for intersection. The number of proximal sj's that are checked for intersection is a variable that can be initialized in the system.
(2) Convex hull intersection test: If di,j becomes less than a predefined threshold value (e.g., the minimum dimension of si or sj), an intersection check between convex hulls hi and hj is performed as a second-level intersection check. If hi∩hj≠NULL, there is a possible intersection between si and sj. (Note that if desired, to reduce later computational burden and increase speed, the convex hull of an element can be precomputed and stored as part of its node information.)
(3) Boundary intersection test: To confirm intersections between si and sj, an intersection test (test for si∩sj) between the boundary representations of si and sj is required—a calculation which must be performed by the solid modeler. This is generally computationally expensive, having time O(n2) where n is the nominal number of entities (vertices, edges and faces) in s1 (or in sj).
Thus, distance checks based on Pi,j's reduce the computationally expensive si∩sj determination by performing it only when necessary. In effect, the ranking of Pi,j's translates geometric proximity between elements to topological proximity in the Shape Modeling Graph S. This potentially reduces the number of entities to be traversed and/or searched, with the additional advantage of being faster than corresponding checks in geometric space.
In addition to using the foregoing method of intersection detection to reduce computing time, an alternative or additional heuristic that may be utilized to reduce the search space for potential intersections is based on the observation that given three entities si, sj and sk (si being the active entity which is being added/subtracted/edited), if sj is closer to si than sk in the topological graph space (i.e., in S), it is in general also closer to si in the geometric space. This means that an intersection is more likely to occur between entities that are closer in the topological space of S than between entities that are further apart in S. Therefore, if sj is closer to si than sk in S, sj should be checked for intersection before sk.
Based on this topological distance heuristic, the shape elements in S can be grouped into the following three categories:
Proximity Category 1: All sj (j≠i, j=1 . . . N) whose parent attachment face is the same as the parent attachment face of si (i.e., same parent, same parent face);
Proximity Category 2: All sj (j≠i, j=1 . . . N) not in Proximity Category 1 whose parent is the same as the parent of si (i.e., same parent, different parent face); and
Proximity Category 3: All sj (j≠i, j=1 . . . N) that do not belong to Proximity Category 1 or Proximity Category 2 (i.e., different parent).
This results in every Pi,j being placed in one of these three proximity categories, which are denoted Ps1 i,j, Ps2 i,j, and Ps3 i,j, with the numbers in the superscripts referring to their respective proximity categories. (The “5” set forth in each superscript refers to the proximity links being soft, the meaning of which will be discussed later.) An illustration is shown in FIG. 33, wherein r6 (i=6, si=r 6) is being edited and the various Pj,6's (j≠6) for r6 are shown. Since the distance between r6 and b2 is zero, P2,6 contains a distance of zero (and hence is not shown). FIG. 33 shows a Category 1 proximity link (Ps1 5,6) ) between r5 and r6, since r5 and r6 share the same parent face; a Category 2 proximity link (Ps2 3,6) between r3 and r6, since r3 and r6 share the same parent (but not the same parent face); and two Category 3 links, the first between r4 and r6 (Ps3 4,6) and the second between bi and r6 (Ps3 1,6).
As illustrated by this example, the following heuristic can be inferred: when si is edited, Proximity Category 1 intersections are most likely followed by those in Proximity Category 2, and subsequently by Proximity Category 3 intersections. To explain another way, elements that share a parent face are more likely to intersect each other than elements on other faces, and elements lying on a single element are more likely to intersect than elements that do not.
Another illustration of this heuristic is shown in FIG. 34A, where element b5 is being edited. FIG. 34B shows the P links for b5 in S at this time step. When b5 is interactively relocated, it is necessary to determine if it intersects any other elements in the model at any time step. To do this, Proximity Category 1 Pj,5's (j=4, 6, 7, 8) from the previous time-step (in this case, P4,5, P6,5, P7,5, P8,5) are first checked for intersection. This is followed by checking Proximity Category 2 links (P2,5, P3,5) for intersection, and then Proximity Category 3 links (P9,5) for intersection (assuming that time is available). The intersection tests are preferably done in the manner previously noted, i.e., within each Proximity Category, convex hull distance checks are done for the Category's elements in ranked order, etc. It is possible that in some cases, some intersections may not be detected because insufficient time is available; however, by performing the distance checks in the aforementioned manner, the probability of missed intersections is greatly decreased.
As previously noted, the purpose of the proximity links is to store the minimum distance between si and all sj (j=1 . . . N, j≠i). This distance is used to create a sequence of sj in increasing order of di,j (j=1 . . . N, j≠i) within each Proximity Category, this sequence being stored in si. These proximity links may be referred to as soft proximity links (Ps i,j) because they enforce system-generated constraints (such as si and sj not to intersect, i≠j) which can be over-ridden by the designer during interactive editing. The modifier soft is used because the constraints can be overridden.
Another type of proximity link is the hard proximity link, Ph i,j. This is a special type of soft proximity link Ps i,j that incorporates a distance-based constraint that must always be satisfied (hence the modifier hard). This distance-based constraint (which could be a minimum distance constraint or a maximum distance constraint) is typically specified by the designer in D in the form of an explicit design rule (as previously discussed in relation to the Design Editing Layer). This rule applies a constraint to the distance between si and sj (i≠j) and is stored in the Ph i,j link. To illustrate, in FIG. 35, the hard proximity link Ph1 3,4 stores the constraint that ribs r3 and r4 are to be kept at a minimum distance x* from each other. This constraint is specified via a design rule in D (shown as R3,4). During interactive editing, if the distance between the ribs (x) becomes less than x* (i.e., if x<x*), this violation of the rule is detected by Ph1 3,4 and the designer is warned about an impending design violation. Finally, for completeness, the corresponding soft link (Ps1 2,3) between b2 and r3 is also shown.
While Ps i,j enables detection of intersection between si and sj, once an intersection is determined to have occurred, appropriate action needs to be taken, i.e., D and S must be updated to reflect the modified geometry. Depending upon the Proximity Category of Pi,j between si and sj, the intersection can be of two types:
(1) Ia, an intra-face intersection between si and sj (i.e., si and sj are linked by a Ps1), or
(2) Ir, an inter-face intersection between si and sj (i.e., si and sj are linked by a Ps2 or Ps3).
As stated earlier, once an intersection (of type Ia or Ir) is detected, the designer is warned before being allowed to proceed because there exists a potential violation of the design intent. If the designer decides to proceed with editing the shape, the D and S are appropriately modified as follows.
For an Ia intersection between si and sj, the modification of S is based on a procedure called splitting, which decomposes si and sj into disjoint sub-elements. This is illustrated in FIG. 36A, where b2 is interactively relocated so as to intersect b3. Once the intersection is detected, b2 and b3 are decomposed into five smaller block elements (b4 through b8) as shown in FIG. 36B (the corresponding S is also shown). The resulting five disjoint elements in S are treated as new block elements, and the original b2 and b3 (in S) are discarded. This is because the intersection of b2 and b3 requires that S be recomputed for accurate graphical display, and since the original b2 and b3 are insufficient for this purpose, the model itself is updated.
The result of splitting changes the nature of the originally added geometry, which therefore requires that D be modified. The designer must select from one of the following choices as being the resultant D (depending on the desired intent):
(1) Five separate elements (i.e., the result of splitting—b4 through b8);
(2) The original elements—b2 and b3; and
(3) A single “new” element containing all the elements b4 through b8.
These interpretations and the corresponding D's are shown in FIGS. 37 through 39 respectively. While the first interpretation is similar to the representation of S, i.e. five separate sub-elements, the second and third interpretations are quite different. The second interpretation, which utilizes the two original entities, treats the two elements independently of their intersection. The third interpretation requires a procedure called grouping which allows the designer to treat the resultant entity as a single entity but only in D, i.e., S remains unchanged from the results of the splitting procedure described earlier. This illustrates the fact that regardless of the interpretation by the designer, S is always the same. The mapping that exists between each Fi in D to one or more sj's in S allows corresponding modification of the nodes in S depending on the interpretation chosen:
Interpretation 1: D.b4maps to S.b4; D.b5 maps to S.b5; and soon.
Interpretation 2: D.b3maps to S.b6+S.b8+S.b7; D.b4maps to S.b4+S.b8+S.b5.
Interpretation 3: D.g1 maps to S.b4+S.b5+S.b6+S.b7+S.b8.
As opposed to the Ia intersection, which results in new geometric entities being created, Ir intersections typically result in the formation of new A links from P links. This is illustrated in FIG. 40A, which shows two examples of interactive editing of a shape resulting in the same local topology for c7. The basic topology of the shape (prior to editing, and showing only Ap 5,7) is illustrated in FIG. 40B.
In Case I of FIG. 40A, the length of cylinder c7 is modified so that it finally intersects b6. In Case II of FIG. 40A, the same resultant topology for c7 is obtained by moving block b2 as illustrated until cylinder c7 finally intersects b6. In either case, when c7 is edited so that c7 approaches b6, Ps3 6,7 is updated and moved up in ranking based on d6,7(S being shown in FIG. 40C). If the distance between c7 and b6 becomes zero (P6,7 contains d6,7=0), a warning message is provided to the designer about the intersection of c7and b6. If the designer allows the intersection to occur, Ps3 6,7 (shown in FIG. 40D) is updated to become a secondary attachment link As 6,7 (as in FIG. 40E) since now there exists contact between b6 and c7. This link is secondary because the original intent of the designer was to add c7 to b5 only, and not to b6.
A geometry link G stores pre-computed geometric information required for interactive modeling. Such information is computed for each si, or between si and sj (i≠j). Pre-computation of geometry allows high-speed interactive editing by virtue of avoiding the need to perform computationally expensive calculations at the time of editing. Examples of such computations include opposite face determination (e.g., for addition of through-holes), or parallel face determination (e.g., for the addition of bridge elements between two elements). Examples of opposite and parallel face geometry links follow; however, it should be understood that other types of geometry links (e.g., between other types of surfaces) can also be defined based on the need of the application.
FIG. 41 shows an “opposite face” geometry link Go (the superscript “o” indicating that the link is for an opposite face relationship) for node b1 of S. The link Go 1,1 relates b1.f1 and b1.f6 as opposite faces. Similar links are constructed for all pairs of faces in b1 (though these are not illustrated in FIG. 41 for sake of clarity). Thus, all Go links start and end on a single node. Pre-computation of opposite faces is useful because in general, when a through-hole is added to an entity, its most common orientation is perpendicular to the face (and thus intersects the opposite face). Pre-computing the opposite face relation and storing it, rather than performing the opposite face determination during editing, enables a higher degree of interactivity (particularly in a time-critical VE environment).
FIG. 42 then illustrates a “parallel face” geometry link Gp (the superscript “p” indicating that the link is for a parallel relationship between the faces of two distinct elements), more particularly, Gp 2,3 illustrates that b2.f4 is parallel to b3.f5. This information is precomputed and the geometry link Gp 2,3 is stored between b2.f4 and b3. f5 for use during subsequent editing. An example of such an edit is shown in FIG. 43, wherein d4 is added to b2.f4 and is then extended until it touches b3.f5 to form a bridge element. Comparing FIGS. 42 and 43 for such a construction, the geometry link with the parallel face allows determination of the affected face of b3 (i.e., the face of b3 to be intersected by d4) in a single topological step.
To summarize, the G link is a specific type of geometric construct that focuses on identifying a specific relationship between si.fp and si.fq (p≠q) (i.e., between different faces on the same element) or between si.fp and sj.fq (i≠j) (i.e., between different faces on different elements). The G link is provided because it is found that relevant relationships, e.g., opposing faces on the same element or parallel faces on different elements, often have practical significance in real-world designs. As another example, consider the design shown in FIG. 44A, wherein it is desired that a cylindrical through-hole pass through r2, r3 and r4 at the location shown. Upon addition of the through-hole to r2.f1, Go 2,2 can be used to determine the next face (r2.f2) that the hole intersects. Next, the geometry link Gp 2,3 may be used to determine the face of r3 (r3. f1) on which the hole continues, and so on. Thus, the presence of G links allows the rapid computation of necessary intersections for the addition of a through-hole by restricting the faces for which intersection calculations must be performed.
As previously noted, VDSF is presently implemented in a three-dimensional VE (i.e., it is a VR-CAD system), though it could be implemented in a 2D system as well. Hand motion input is implemented through the Fifth Glove™ glove and the Ascension Technology Corporation Flock of Birds™ six degrees of freedom tracking devices with the Xtensory Xvs-SyncLink™ library. However, VDSF may use a wide variety of alternative or additional motion input systems. Six degree of freedom (6DOF) motion input devices may be generally categorized as tracking devices, point input devices, and glove-type devices. Tracking devices, which may operate based on mechanical, ultrasonic and magnetic optical principles, are used in tracking the user's hand and or head motion and orientation. Point input devices include the 6DOF mouse and the force ball. A 6DOF mouse, e.g., the Flying Mouse™ from Logitech, functions like a normal mouse on the desktop, but as a six-degree-of-freedom device, it allows the user to provide height input once the mouse is lifted from the desktop. A force ball, e.g., the Spaceball™ from Spaceball. Technology, Inc., uses mechanical strain experienced by a gripped ball to measure the forces and torques that the user applies in each of the possible directions. Glove-type devices, e.g., the CyberGlove™ from Virtual Technologies, Inc., consist of a wired cloth that is worn over the hand like a normal glove. Fiber-optical, electrical, or resistive sensors are used to measure the positions of the joints in the finger so that gesture input may be obtained.
Voice command input has been implemented using the Command Corp, Inc. In-Cubed™ voice recognition system. Other systems having suitable voice recognition capabilities, e.g., Voice Assist from Soundblaster and Dragon Dictate from Dragon Systems, Inc., could also be useful in the present invention.
Three-dimensional visual design output is provided to each user via a pair of Stereographics Crystal Eyes™ LCD shutter glasses. The rendered design is shown on a 60-inch inclined screen (intended to emulate a drafting table) upon which stereoscopic images are projected via a stereo-capable video projector. Rather than using this mode of display, VDSF could instead use a head-mounted display (HMD) wherein two display screens show independent views for each eye, thereby creating a stereoscopic image. Examples of HMDs are Virtual-I/O I-Glasses™ and MRG2™ from Liquid Image Corporation, VR4™ from Virtual Research Systems, and General Realities Cyber Eyes™. However, the use of shutter glasses and screen-displayed stereo images is preferred because this more conveniently accommodates simultaneous use by multiple users.
Auditory output (when used) is provided by the Soundblaster™ AWE-32 sound card. It is possible to provide more complex auditory output wherein sounds appear to come from different locations in the VE; such a system is provided by Alphatron™ from Crystal River Engineering.
Tactile output can be accommodated by VDSF, but has not been found to be particularly useful when provided by the common tactile output devices found on the market in 1999. Tactile feedback allows a user to touch and feel the resistance of objects in the VE by applying electrical signals or vibrations to different points on the user's hand, by inflating bladders installed within a glove or similar device, or by using mechanical devices to exert force on the user. Examples of these systems are the Cybertouch™ tactile feedback system from Virtual Technologies, Inc., and the Sensible Devices, Inc. Phantom™ and Exos, Inc. Safire™ force-reflecting interfaces.
Visual output is implemented through the Open-GL™ library on the Silicon Graphics™ Indigo (SGI) workstation, which is networked to a Pentium Pro™ 200MHz personal computer. For graphics/display, VDSF uses WorldToofKit™ (WTK), which is a library of “C” or “C++” functions that allows interfacing with a wide range of input and output. The geometric modeling and interaction functions have been implemented utilizing WTK. The entire system is interfaced to the ProEngineer™ CAD system to allow further modification of VDSF designs, although other CAD systems could be used instead.
Further VDSF hardware and software details are provided in TABLE 2. The architecture of the prototypical system and its hardware and software components are shown in FIG. 45.
VDSF HARDWARE AND SOFTWARE
Custom Frame (with mirror,
Pentium ™ Pro PC:
Ascension ™ trackers and emitter
WorldToolKit ™ (Sense8 Corp.)
Pair of 5th ™ Gloves
Microsoft Visual C++ ™
Crystal Eyes ™ LCD glasses
IBM Voice Type ™
The current functionality of VDSF is summarized in TABLE 3.
SUMMARY OF VDSF FUNCTIONALITY
Commands (via WISL interface)
Functions for any activity that excludes shape editing
Allows viewing the desired location on the design.
(i) Flying: follow 3D locator.
(ii) Zoom In/Out: zoom in or out in current
(iii) Rotate viewpoint: rotate the viewpoint
Selection allows entities in the model to be
selected. Two modes:-
(i) Use 3D locator and issue spoken “Select” command
(ii) Use spoken command to step through all entities
until the desired entity is highlighted
Allows grasping and manipulation of design
to get the desired view
(i) Designer uses hand input and tracker to grab and
Functions for shape editing
Pre-defined entities: block, rib, cylinder, cone,
wedge, etc. (Positive or negative entities)
Hand input to indicate location, and spoken
commands (e.g., “Block”) to instance the
shape to be added to that location
Select entity and issue “remove” spoken command
Select entity and issue “delete” spoken command
(different from “remove,” which detaches a feature
but does not delete it from the model)
Via spoken commands such as “Increase *”
(where * is a parameter for the element in
question, such as “length”, “width”,
“height” for block element)
Interactive sizing using 3D hand input
or spoken command (e.g., “Scale”) until
desired size is reached
(i) Change x, y, z location via
(ii) Constrained translation (via hand input)
Grab and rotate entity
Some examples of shapes modeled by VDSF are shown in FIGS. 32 through 53 to illustrate the effects of VDSF's dual graph representation:
The object in FIG. 46 shows a simplified shape of a telephone component designed in VDSF. It consists of four block elements (b1, b2, b3, and b4) and an element p5 that represents the numeric pad (composed of a base and several blocks representing buttons, these elements being denoted with a superscript “p”). The D and S for the pad p5 are shown in FIG. 47, and the D and S for the overall phone model (wherein p5 is considered as a single component) are shown in FIG. 48. Block b1 is the base (parent) to which the other components—antenna (b2), earpiece (b3), mouthpiece (b4) and buttons (p5)—are attached. FIG. 48 shows D for the model in FIG. 46; FIG. 49 shows S (with only the A links illustrated); and FIGS. 50 and 51 respectively show the P and G links in S.
FIG. 52A illustrates an injection-molded part (Part I) designed in VDSF. The corresponding D and S are shown in FIGS. 52B and C respectively (note that P and G links are not shown for sake of clarity).
FIG. 53A illustrates another injection-molded part (Part II) designed in VDSF. The corresponding D and S are shown in FIGS. 53B and C respectively (again note that P and G links are not shown for sake of clarity).
FIGS. 54A-54C illustrates a machined part constructed in VDSF and its corresponding D and S (with only the relevant links in S being shown). FIGS. 54A-54C illustrates the use of a special function added to VDSF which might be called the “Add In” operation (as opposed to the “Add On” operation) in that it allows insertion of one element into another (in this case c8 into b4).
FIGS. 55A-55C illustrate another machined part constructed in VDSF and its corresponding D and S. Again, only the relevant links in S are shown.
FIGS. 56A and 56B illustrate the successive stages of construction of an injection-molded part (Part V) in VDSF; D and S are not shown.
To review, the notion behind use of the dual D- and abased representation is to maintain a clear separation between design intent and geometry for fast display. D provides a higher-level hierarchical representation suitable for storing design intent, whereas S is a more detailed representation which preserves this hierarchy and in addition uses a variety of links (i.e., attachment A, proximity P and geometry G) to enable high-speed interactive editing. A representation scheme such as the one used by VDSF can be evaluated for several characteristics (see, e.g., M. Mantyla, “An Introduction to Solid Modeling,” Computer Science Press):
1. Expressive power: The types of shapes that can be modeled with S include most manufacturable mechanical shapes, e.g., those that are commonly found in injection molding, casting, forging, and machining.
2. Validity: The designer creates a shape by adding elements (positive or negative) via the “Add On” operation. The constraints on this operation prevent the generation of an invalid model, and the resultant S must always represent a valid model.
3. Unambiguity and uniqueness: S is not unique since more than one S can represent a given geometrical shape (examples of which were given earlier).
COMPARISON OF CHARACTERISTICS OF D AND S GRAPHS
Allow rapid display
and interaction with
design (restricted to
(based on D)
Types of links:
Represent holes and
(Soft & Hard)
For complexity analysis, the following quantities are assumed:
Number of shape elements in S.
Average number of faces in each element si.
Average number of edges in each element si, assumed = O(M).
Average number of faces in each element si, assumed = O(M).
Constant time operation.
It is known from previous work that an intersection determination between the convex hulls hi and hj of si and sj can be performed in O(ni+nj)=O(M) (where ni and nj are the number of vertices in the linearized model of si and sj respectively). (See, e.g., D. Dobbin and D. G. Kirpatrick, “A Linear Algorithm for Determining the Separation of Convex Hull Polyhedra,” Journal of Algorithms, vol. 6, pp.381-392.) Such convex hull intersection checks are used in the P link computation previously discussed.
The fundamental operations in D and S are Add Entity, Edit Entity and Delete Entity, which have the following complexity (the results for the following complexity analysis being summarized in TABLE 5, it being noted that some results for D were already reported in TABLE 1):
1) Add a link (or links) with negative/positive:
2) Add links for design rules:
1) Add link (or links) that map D to S
A5 link (if negative entity)
2) Compute auxiliary links
P5 links (with convex hull)
P6 links (if exact)
Sort P links
Categorize P links (into Category 1, 2, 3)
Faces with hole
Faces without hole
Thus, for element addition, the best case performance is O(NM) and worst case O(NM2). It is noted that a typical solid modeler is O(NM2)). However, from the above description, it is seen that the principal reason for the O(NM2) term comes from the Gp links and Ps (with exact computation). This computation can potentially be reduced by using heuristics based on the knowledge of the entity type.
An editing operation could be of two types: edit location and edit size. In both cases, the computational requirements are similar and are discussed below.
1) Update parameter values
1) Check Proximity Category 1 distance
If intersection occurs:
Best case (two entities intersect)
Worst case (all entities intersect)
2) Repeat 1) for Category 2 and 3 if time is available
[Same as above]
3) Update P's
Calculate minimum distances
with convex hull based
Categorize and sort P
Thus, while editing, best case performance is O(1) (if no intersection), and worst case is O(NM2) (similar to a solid modeler). On average, computations will lie in between these two extreme cases.
Note that in general, addition of an entity is more computationally expensive than editing an entity. This is in consonance with the philosophy of pre-calculating several properties to allow faster editing later.
1) Delete entity from D
1) Remove Ap and As
2) Remove P links
3) Update resultant P links
4) Remove G links
5) Delete children recursively
RESULTS OF COMPLEXITY ANALYSIS
VDSF was designed as a stand-alone system whereby prototype models could be rapidly designed in a VR-CAD environment, with the intent that the designed models could then be easily transferred to any number of conventional CAD systems for detailed design (i.e., finalization). One approach used by VDSF is to use the ProDevelop™ API (Application Programming Interface) in the ProEngineer CAD environment to asynchronously create a parallel representation of VDSF's VR-CAD model in the conventional CAD environment. Once the design is completed in VDSF, the parallel CAD model can be used for further detailed design in a conventional CAD system. A second approach is to have VDSF output a data file in a format suitable for reading by the desired commercial CAD system. The designs created by use of VDSF are therefore usable in all major and well-known commercial CAD systems.
Within the preferred version of the invention described above (VDSF), the Virtual Modeler Layer implements the dual graph representation, and the user creates designs in a VE environment by use of voice and motion input via the Interaction Mechanisms Layer. However, it is noted that the advantages of the dual graph representation may be realized in conventional 2D-CAD systems as well as in VR-CAD systems utilizing a VE. The interactivity advantages of the dual graph representation will probably find their greatest value in VE applications owing to the time-sensitive nature of these systems, but these advantages are nevertheless still useful in 2D applications.
While it was previously noted that certain constraints are provided to allow faster design—e.g., the constraint that addition of elements will situate them in face-to-face abutment—it should be understood that these constraints are not strictly necessary, and other (or no) constraints might be used. For example, consider the shape in FIG. 40A. Creating this shape via the “Add On” method would require the designer to create two block elements b2 and b3 and then “add” them to b1 (FIGS. 58A and B). An easier way to create this shape is to add a block b4 directly to b1 as shown in FIGS. 59A, 59B and 59C via the “Add In” operation (i.e., the aforementioned constraint could be relaxed to allow intersection/insertion of elements rather than only abutment of elements). Depending on the specific geometry to be created, the “Add In” mode might allow faster creation as compared to the “Add On” mode. In other words, depending on the context of the design effort, different constraints may be appropriate.
As another example of effects that can be attained when constraints are relaxed or eliminated, consider that VDSF could eliminate the requirement for face-to-face element attachment (i.e., face-to-edge contact as in FIG. 9C, or face-to-vertex contact, could be allowed). If this was done, the complexity of the dual graph representation would likely need to increase in order for it to support the newly-available design possibilities; for example, attachment links would need to be added for element edges and/or vertices as well as for element faces. Additionally, some operations currently supported by D and S may grow in complexity to such an extent that they may no longer be feasible; for example, the calculation of orthogonal bounding boxes may be too burdensome to support the desired degree of interactivity. The implicit design constraints noted earlier in this document are highly advantageous in that they support efficient use of the dual graph representation, but they are not necessarily required for its use.
|Cited Patent||Filing date||Publication date||Applicant||Title|
|US5177689||Mar 13, 1990||Jan 5, 1993||Mitsubishi Denki K.K.||Cad/cam apparatus for enhancing operator entry of machining attributes and geometric shapes|
|US5504854||May 18, 1994||Apr 2, 1996||Fuji Xerox Co., Ltd.||Apparatus and method for inputting cell shape and position and inter-cell calculation|
|US5659493||Mar 3, 1995||Aug 19, 1997||Ford Motor Company||Virtual machining techniques for modifying computer models of parts|
|US5742291||May 9, 1995||Apr 21, 1998||Synthonics Incorporated||Method and apparatus for creation of three-dimensional wire frames|
|US5758122||Mar 16, 1995||May 26, 1998||The United States Of America As Represented By The Secretary Of The Navy||Immersive visual programming system|
|US5923573 *||May 30, 1997||Jul 13, 1999||Honda Giken Kogyo Kabushikikaisha||Three-dimensional CAD system for producing a three-dimensional model|
|USRE36602 *||Jun 26, 1997||Mar 7, 2000||The Trustees Of The Stevens Institute Of Technology||Concurrent engineering design tool and method|
|1||*||"Solid Modeling" by Vadim Shapiro, Technical Report SAL 2001-2, Oct. 2001.*|
|2||A. Stork, M. Maidhof, "Efficient and Precise Solid Modeling using a 3D Input Device," Proc. of the Fourth Symposium on Solid Modeling & Applications, Atlanta, GA, May 14-16, 1997, pp. 181-194.|
|3||A.G. Requicha, "Representation for Rigid Solids: Theory Methods & Systems,", Computer Survey, v12, No. 4, Dec. 1990, pp 438-464.|
|4||*||AutoCAD and its applications Advanced Release 14, by Terence M. Shumaker et al., Autodesk Press, ISBN: 1-56637-414-6, May, 1998, pp. 219, 220, and 514.*|
|5||C.C. Chu, T.H. Dani and R. Gadh, "Multimodal interface for a Virtual Reality Based CAD System", Proc. 1997 IEEE Intl. Conf. on Robotics & Automation, vol. 2, pp. 1329-1334.|
|6||D. Dietz, "Real Engineering in a Virtual World," Mechanical iEngineering Magazine, vol. 117, No. 7, Jul., 1995, pp. 78-85.|
|7||D. Weimer, S.K. Ganapathy, "A Synthetic Visual Environment with Hand Gesture and Voice Input," Proceedings of the ACM Conference on Computer Human Interfaces, May 1989, pp. 235-240.|
|8||D.L. Jenkins, R.R. Martin, the importance of free-hand sketching in conceptual design: Automatic Sketch input,: Process of the 1993 ASME Design Technical Conference, Albuquerque, NM, Sep. 19-22, 1993, pp. 115-128.|
|9||D.P. Mapes, J.M. Mosehll, "A two-handed Inerface for Object Manipulation in a Virtual Environment," Presence, vol. 4, No. 4, Fall 1995, pp. 403-416.|
|10||*||Danial T. Banach, Mechanical Desktop 2.0, International Thompson, Copyright 1998, pp. 82-107 and 214-257.*|
|11||*||Danial T. Banach, Mechanical Desktop 2.0, International Thompson, Copyright 1998.*|
|12||E. Herranz, R.A. Bolt, "Two-Handed Gesture In Multi-Modal Natural Dialogue", Proceedings of the ACM Symposium on User Interface Software Technology, Nov. 15-18, 1992, pp. 7-14.|
|13||E. Sachs, A. Roberts, D. Stoops, "3-Draw: A Tool for Designing 3D Shapes," IEEE Computer Graphics & Applications, vol. 11, No. 11, pp. 18-26, 1991.|
|14||G. Wyvill, T.L. Kunii, "A functional model for constructve solid geometry", The Visual Computer, v. 1, 1985, pp. 3-14.|
|15||J. Liang, "JDCAD: A Highly Interactive 3D Modeling System," Computer Graphics, v. 18, No. 4, 1994, pp. 499-506.|
|16||J.B. Allan, B. Wyvill, I.H. Witten, "A Methodology for Direct Manipulation of Polygon Meshes," New Advances in Computer Graphics: Proceedings of CD International '89, pp. 451-469.|
|17||K. Lee, D.C. Gossard, "A hierarchical data structure for representing assemblies: part 1," CAD Journal, v. 17, No. 1, Jan./Feb. 1985.|
|18||M. Fa, T. Fernando; and Dew, P.M. 1993. "Interactive Constraint-based Solid Modeling using Allowable Motion", Proc. of the 2<nd >Symposium on Solid Modeling and Applications, 242-252.|
|19||M. Fa, T. Fernando; and Dew, P.M. 1993. "Interactive Constraint-based Solid Modeling using Allowable Motion", Proc. of the 2nd Symposium on Solid Modeling and Applications, 242-252.|
|20||M. Figueiredo, K. Bohm, J. Teixeira, "Advanced Interaction Techniques in virtual Environments," Computers and Graphics, v. 17, No. 6, pp. 655-661, 1991.|
|21||*||Mechanical Desktop 2.0 Applying Designer and Assembly Modules, by Daniel T. Banach, Autodesk Press, ISBN: 0-7668-0068-7, Jun. 1998, pp. xv, 3, 8, 186, and 224.*|
|22||P. Zhigeng, Z. Mingmin, Z. Wenting, S. Jiaoying, "Time-Critical Computing in Virtual Environments," Proc. of the 4<th >International Conference on Computer-Aided Design and Computer Graphics, Oct. 23-25, 1995, Wuhan, China, pp. 712-717.|
|23||P. Zhigeng, Z. Mingmin, Z. Wenting, S. Jiaoying, "Time-Critical Computing in Virtual Environments," Proc. of the 4th International Conference on Computer-Aided Design and Computer Graphics, Oct. 23-25, 1995, Wuhan, China, pp. 712-717.|
|24||P.M. Hubbard, "Approximating Polyhedra with Spheres for Time-Critical Collision Detection," ACM Transactions on Graphics, v. 15, No. 3, Jul. 1996, pp. 179-201.|
|25||P.R. Wilson, M.J. Pratt, "A taxonomy of features for solid modeling," Geometric Modeling for CAD Applications, M.J. Wozny, H.W. McLaughlin, J.L. Encarnacao (Eds.), Elsevier Science Publishers, 1988, pp. 125-137.|
|26||R.A. Bolt, "Put That There: Voice and gesture at the graphics interface," Computer Graphics, vol. 14, No. 2, pp. 262, 1980.|
|27||T. Yoshimura, Y. Nakamura, M. Suguira, "3D Direct Manipulation Interface: development of Zashiki-Warashi System," Computers & Graphics, vol. 18, No. 2, pp. 201-207, 1992.|
|28||T.H. Dani and R. Gadh, "Virtual Reality: A New Tool for the Mechanical Engineer," Mechanical Engineers Handbook, Myer Kutz, (Ed.), John Wiley & Sons, 1997.|
|29||T.H. Dani, R. Gadh, "Creation of Concept Shape Designs via a Virtual Reality Interface," CAD Journal, Elsevier Publications, 29(7), 1997.|
|30||Trika and Banerjee and Kashyap, "Virtual Reality Interface for Feature Based Computer Aided Design Systems", 1997.|
|31||Y.A. Tijerno, K. Mochizuiki, F. Kishino, "Interactive 3-D Computer Graphics Driven Through Verbal Instructions: Previous and Current Activities at ATR," Computer & Graphics, vol. 18, No. 5, pp. 621-631, 1994.|
|Citing Patent||Filing date||Publication date||Applicant||Title|
|US6839606||Oct 24, 2001||Jan 4, 2005||Delphi Technologies, Inc.||Horizontally structured manufacturing process modeling for fixtures and tooling|
|US6859201 *||Jan 28, 1999||Feb 22, 2005||Autodesk, Inc.||Method and apparatus for viewing a mechanical design|
|US6882347 *||Nov 30, 2000||Apr 19, 2005||Microsoft Corporation||Texture generating apparatus for dynamic interference checking|
|US6941530 *||Mar 24, 2003||Sep 6, 2005||Texas Instruments Incorporated||Method of cross-mapping integrated circuit design formats|
|US6950719||Jan 31, 2003||Sep 27, 2005||Delphi Technologies, Inc.||Horizontally structured manufacturing process modeling: across file feature operability|
|US6964023 *||Feb 5, 2001||Nov 8, 2005||International Business Machines Corporation||System and method for multi-modal focus detection, referential ambiguity resolution and mood classification using multi-modal input|
|US6985793||Jan 31, 2003||Jan 10, 2006||Delphi Technologies, Inc.||Horizontally structured CAD/CAM coordinate system for manufacturing design|
|US7024272||Feb 5, 2003||Apr 4, 2006||Delphi Technologies, Inc.||Virtual design, inspect and grind optimization process|
|US7043695 *||Jun 25, 2001||May 9, 2006||Technion Research & Development Foundation Ltd.||Object positioning and display in virtual environments|
|US7079908 *||Jan 31, 2003||Jul 18, 2006||Delphi Technologies,Inc.||Horizontally-structured modeling for analysis|
|US7079990 *||Feb 8, 2001||Jul 18, 2006||Solidworks Corporation||Automated connections of computer-aided design components|
|US7086028 *||Apr 9, 2003||Aug 1, 2006||Autodesk, Inc.||Simplified generation of design change information on a drawing in a computer aided design (CAD) environment|
|US7099803 *||Sep 6, 2000||Aug 29, 2006||Proficiency Solutions Ltd.||Data exchange between computer aided design systems|
|US7099804||Oct 24, 2001||Aug 29, 2006||Delphi Technologies, Inc.||Automated horizontally structured manufacturing process design modeling|
|US7110849||Oct 24, 2001||Sep 19, 2006||Delphi Technologies, Inc.||Horizontally-structured CAD/CAM modeling for virtual fixture and tooling processes|
|US7155375 *||Feb 8, 2002||Dec 26, 2006||Impactxoft||Method and system for designing objects using design intent merge|
|US7176942 *||Jun 22, 2001||Feb 13, 2007||Dassault Systemes||Collaborative design|
|US7191104 *||Jul 11, 2002||Mar 13, 2007||Ford Global Technologies, Llc||Method of real-time collision detection between solid geometric models|
|US7239985 *||Sep 23, 2003||Jul 3, 2007||Ncr Corporation||Methods, systems, and data structures for modeling information quality and maturity|
|US7245984||Jan 31, 2003||Jul 17, 2007||Delphi Technologies, Inc.||Horizontally structured manufacturing process modeling: exterior linked representational embodiment|
|US7280948||May 10, 2002||Oct 9, 2007||Delphi Technologies, Inc.||System and method for integrating geometric models|
|US7283136 *||Mar 23, 2001||Oct 16, 2007||Dassault Systemes||Cell descriptor|
|US7308386||Oct 24, 2001||Dec 11, 2007||Delphi Technologies, Inc.||Enhancement to horizontally-structured CAD/CAM modeling|
|US7313504 *||Oct 11, 2002||Dec 25, 2007||Solidworks Corporation||Model management technology using grouping of features|
|US7421363||Feb 5, 2003||Sep 2, 2008||Delphi Technologies, Inc.||Method for virtual inspection of virtually machined parts|
|US7466313 *||Dec 22, 2000||Dec 16, 2008||Hewlett-Packard Development Company, L.P.||Method for interactive construction of virtual 3D models|
|US7472044||Mar 22, 2004||Dec 30, 2008||Delphi Technologies, Inc.||Horizontally structured CAD/CAM modeling-vertical to horizontal conversion|
|US7483919||Dec 7, 2004||Jan 27, 2009||Almen Laboratories||Object based image retrieval|
|US7492364 *||Jul 23, 2003||Feb 17, 2009||Imagecom, Inc.||System and method for creating and updating a three-dimensional model and creating a related neutral file format|
|US7495662 *||Jul 23, 2004||Feb 24, 2009||Dassault Systemes||Part design system using a context-free grammar|
|US7519603 *||Nov 27, 2002||Apr 14, 2009||Zyvex Labs, Llc||Efficient data structure|
|US7530034 *||Feb 27, 2006||May 5, 2009||Dcg Systems, Inc.||Apparatus and method for circuit operation definition|
|US7568171 *||Jul 23, 2007||Jul 28, 2009||Pacific Data Images Llc||Stroke-based posing of three-dimensional models|
|US7571079||Jan 25, 2006||Aug 4, 2009||Dassault Systems Solidworks Corporation||Aware and active features for computer-aided design systems|
|US7587302||Jan 31, 2006||Sep 8, 2009||Autodesk, Inc.||Graphic interactive method to reorder sequential data values on graphic objects|
|US7625335||Aug 24, 2001||Dec 1, 2009||3Shape Aps||Method and apparatus for three-dimensional optical scanning of interior surfaces|
|US7650260||Feb 8, 2002||Jan 19, 2010||Impact Xoft||Method and system for designing objects using functional object representation|
|US7663625||Dec 6, 2005||Feb 16, 2010||Dassault Systemes||Collaborative design|
|US7755620 *||May 13, 2004||Jul 13, 2010||Interlego Ag||Method and system for manipulating a digital representation of a three-dimensional object|
|US7755621 *||Jul 13, 2010||Autodesk, Inc.||Fast interactive object manipulation|
|US7761266||Jan 25, 2006||Jul 20, 2010||Autodesk, Inc.||Synchronized physical and analytical representations of a CAD model|
|US7773098 *||May 5, 2006||Aug 10, 2010||Canon Kabushiki Kaisha||Virtual reality presentation apparatus and method|
|US7788068||Jan 31, 2006||Aug 31, 2010||Autodesk, Inc.||Transferring load information and result information between analysis and design software|
|US7856342||Jan 3, 2007||Dec 21, 2010||Autodesk, Inc.||Automatic reinforcement modeling|
|US7905610 *||Aug 29, 2006||Mar 15, 2011||Nvidia Corporation||Graphics processor system and associated method for projecting an image onto a three-dimensional object|
|US7908280||Oct 30, 2007||Mar 15, 2011||Nokia Corporation||Query method involving more than one corpus of documents|
|US7917464||Mar 29, 2011||Metacarta, Inc.||Geotext searching and displaying results|
|US7917894||Jun 3, 2005||Mar 29, 2011||International Business Machines Corporation||Constraining source code and objects for analysis tools|
|US7930149 *||Dec 19, 2003||Apr 19, 2011||Sap Aktiengesellschaft||Versioning of elements in a configuration model|
|US7953732||Jun 7, 2005||May 31, 2011||Nokia Corporation||Searching by using spatial document and spatial keyword document indexes|
|US7970587 *||Mar 20, 2008||Jun 28, 2011||Siemens Industry, Inc.||System and method for defining part interfaces|
|US7986318 *||Feb 2, 2006||Jul 26, 2011||Autodesk, Inc.||Volume management system for volumetric displays|
|US7986827 *||Feb 6, 2007||Jul 26, 2011||Siemens Medical Solutions Usa, Inc.||System and method for multiple instance learning for computer aided detection|
|US8015183 *||Jun 12, 2007||Sep 6, 2011||Nokia Corporation||System and methods for providing statstically interesting geographical information based on queries to a geographic search engine|
|US8032337 *||Mar 1, 2002||Oct 4, 2011||3Shape A/S||Method for modeling customized earpieces|
|US8059120 *||Oct 26, 2006||Nov 15, 2011||Fujitsu Limited||Image process apparatus for three-dimensional model|
|US8074182 *||Jul 16, 2008||Dec 6, 2011||Sysmex Corporation||Work procedure display method and system, production process management method and system, and computer program of the same|
|US8098263 *||May 27, 2005||Jan 17, 2012||Canon Kabushiki Kaisha||Image processing method and image processing apparatus|
|US8099260||Jul 27, 2006||Jan 17, 2012||Autodesk, Inc.||Analysis error detection for a CAD model|
|US8135484 *||Aug 8, 2008||Mar 13, 2012||Fujitsu Limited||Work instruction sheet preparing device, method and program|
|US8154545 *||Dec 6, 2005||Apr 10, 2012||Parametric Technology Corporation||Method for producing technical drawings from 3D models with at least two colliding 3D bodies|
|US8200676||Jun 28, 2006||Jun 12, 2012||Nokia Corporation||User interface for geographic search|
|US8201099 *||Sep 26, 2006||Jun 12, 2012||Sandia Corporation||Method and system for rendering and interacting with an adaptable computing environment|
|US8248407 *||Jun 19, 2008||Aug 21, 2012||Dassault Systemes||System and method of computer-aided design of a 3D object modeled by geometries|
|US8291378||Jul 29, 2008||Oct 16, 2012||International Business Machines Corporation||Simplified deployment modeling|
|US8302093||Aug 28, 2008||Oct 30, 2012||International Business Machines Corporation||Automated deployment of defined topology in distributed computing environment|
|US8315840||May 27, 2010||Nov 20, 2012||Autodesk, Inc.||Transferring structural loads and displacements between analysis and design software|
|US8345043 *||Apr 13, 2007||Jan 1, 2013||Autodesk, Inc.||Solving networks of geometric constraints|
|US8346020 *||May 1, 2009||Jan 1, 2013||Zentech, Inc.||Automated generation of 3D models from 2D computer-aided design (CAD) drawings|
|US8402381||Sep 23, 2008||Mar 19, 2013||International Business Machines Corporation||Automatically arranging widgets of a model within a canvas using iterative region based widget relative adjustments|
|US8417658||Sep 12, 2008||Apr 9, 2013||International Business Machines Corporation||Deployment pattern realization with models of computing environments|
|US8473524||Apr 28, 2009||Jun 25, 2013||Dassault Systemes||Method and system for updating object data with respect to object specifications in a product life cycle management system|
|US8510087||Sep 29, 2010||Aug 13, 2013||Siemens Product Lifecycle Management Software Inc.||Variational modeling with discovered interferences|
|US8539005||Feb 24, 2010||Sep 17, 2013||Steer Information Technologies Pvt. Ltd.||Method and system for configuring and visualizing an extruder|
|US8606735 *||Apr 29, 2010||Dec 10, 2013||Samsung Electronics Co., Ltd.||Apparatus and method for predicting user's intention based on multimodal information|
|US8677317||Mar 30, 2012||Mar 18, 2014||International Business Machines Corporation||Simplified deployment modeling|
|US8706450 *||Apr 24, 2008||Apr 22, 2014||Siemens Aktiengesellschaft||Method and system for identification of grouping characteristics|
|US8725466 *||Oct 13, 2009||May 13, 2014||Cct International, Inc.||System and method for hybrid solid and surface modeling for computer-aided design environments|
|US8762877 *||Feb 13, 2009||Jun 24, 2014||Ice Edge Business Solutions Ltd.||Creation and modification of valid functional design layouts|
|US8775451||Aug 20, 2007||Jul 8, 2014||Almen Laboratories, Inc.||Object based image retrieval|
|US8793652||Jun 7, 2012||Jul 29, 2014||International Business Machines Corporation||Designing and cross-configuring software|
|US8799203||Jul 16, 2009||Aug 5, 2014||International Business Machines Corporation||Method and system for encapsulation and re-use of models|
|US8810570 *||Apr 13, 2009||Aug 19, 2014||Siemens Product Lifecycle Management Software Inc.||System and method for active selection in a solid model|
|US8825450 *||Oct 22, 2009||Sep 2, 2014||Dassault Systemes||Method and system for updating a modeled object in a product lifecycle management system|
|US8849987||Jun 30, 2009||Sep 30, 2014||International Business Machines Corporation||Automated discovery of a topology of a distributed computing environment|
|US8872820 *||Apr 13, 2009||Oct 28, 2014||Siemens Product Lifecycle Management Software Inc.||System and method for splitting faces on a solid model|
|US8896597 *||Apr 13, 2009||Nov 25, 2014||Siemens Product Lifecycle Management Software Inc.||System and method for modifying geometric relationships in a solid model|
|US8913059||Nov 29, 2012||Dec 16, 2014||Autodesk, Inc.||Solving networks of geometric constraints|
|US8930415||Aug 10, 2005||Jan 6, 2015||Dassault Systemes||Method and computer system for interacting with a database|
|US8941650 *||Jul 28, 2009||Jan 27, 2015||Autodesk, Inc.||Method for validating features in a direct modeling paradigm|
|US8983803 *||Sep 2, 2011||Mar 17, 2015||Siemens Product Lifecycle Management Software Inc.||Tolerant intersections in graphical models|
|US9015593||Dec 1, 2008||Apr 21, 2015||International Business Machines Corporation||Managing advisories for complex model nodes in a graphical modeling application|
|US9046995 *||Apr 9, 2009||Jun 2, 2015||On24, Inc.||Editing of two dimensional software consumables within a complex three dimensional spatial application and method|
|US9123171 *||Jul 18, 2014||Sep 1, 2015||Zspace, Inc.||Enhancing the coupled zone of a stereoscopic display|
|US9177408||Jun 15, 2011||Nov 3, 2015||Lucasfilm Entertainment Company Ltd.||Modifying an animation having a constraint|
|US9201972||Oct 30, 2007||Dec 1, 2015||Nokia Technologies Oy||Spatial indexing of documents|
|US9223568||Mar 13, 2013||Dec 29, 2015||International Business Machines Corporation||Designing and cross-configuring software|
|US9230360||Sep 10, 2010||Jan 5, 2016||Lego A/S||Connectivity depended geometry optimization for real-time rendering|
|US9235909||May 6, 2008||Jan 12, 2016||International Business Machines Corporation||Simplifying the presentation of a visually complex semantic model within a graphical modeling application|
|US9245381||Jan 31, 2013||Jan 26, 2016||Ice Edge Business Solutions, Ltd||Visual distortion effects through translucent structures in design software|
|US20010007095 *||Dec 22, 2000||Jul 5, 2001||Klaus Kehrle||Method for interactive construction of virtual 3D circuit models|
|US20020033845 *||Jun 25, 2001||Mar 21, 2002||Geomcore Ltd.||Object positioning and display in virtual environments|
|US20020046368 *||Sep 4, 2001||Apr 18, 2002||Wolfgang Friedrich||System for, and method of, situation-relevant asistance to interaction with the aid of augmented-reality technologies|
|US20020063720 *||Nov 30, 2000||May 30, 2002||Williams Ian M.||Texture generating apparatus for dynamic interference checking|
|US20020107673 *||Feb 8, 2001||Aug 8, 2002||Haller Kirk D.||Automated connections of computer-aided design components|
|US20020133252 *||Oct 24, 2001||Sep 19, 2002||Landers Diane M.||Horizontally structured manufacturing process modeling for fixtures and tooling|
|US20020133253 *||Oct 24, 2001||Sep 19, 2002||Landers Diane M.||Horizontally-structured CAD/CAM modeling for virtual fixture and tooling processes|
|US20020133265 *||Oct 24, 2001||Sep 19, 2002||Landers Diane M.||Horizontally structured manufacturing process modeling for concurrent product and process design|
|US20020133266 *||Oct 24, 2001||Sep 19, 2002||Landers Diane M.||Horizontally structured manufacturing process modeling for alternate operations, large parts and charted parts|
|US20020133803 *||Oct 24, 2001||Sep 19, 2002||Landers Diane M.||Enhancement to horizontally-structured CAD/CAM modeling|
|US20020135618 *||Feb 5, 2001||Sep 26, 2002||International Business Machines Corporation||System and method for multi-modal focus detection, referential ambiguity resolution and mood classification using multi-modal input|
|US20020152000 *||Oct 24, 2001||Oct 17, 2002||Landers Diane M.||Automated horizontally structured manufacturing process design modeling|
|US20020180735 *||Mar 23, 2001||Dec 5, 2002||Valentin Chartier||Cell descriptor|
|US20020183878 *||Jun 22, 2001||Dec 5, 2002||Valentin Chartier||Collaborative design|
|US20030085889 *||Oct 11, 2002||May 8, 2003||Ricardo Chin||Model management technology using feature chains|
|US20030130749 *||Nov 7, 2002||Jul 10, 2003||Albert Haag||Multi-purpose configuration model|
|US20030179232 *||Mar 25, 2002||Sep 25, 2003||Fousek Daniel P.||Custom drawing symbols|
|US20030204284 *||Feb 5, 2003||Oct 30, 2003||Thomas Steven M.||Method for virtual inspection of virtually machined parts|
|US20030204285 *||Feb 5, 2003||Oct 30, 2003||Thomas Steven M.||Virtual design, inspect and grind optimization process|
|US20030204286 *||Feb 5, 2003||Oct 30, 2003||Thomas Steven M.||Method for modeling complex, three dimensional tool paths through a workpiece|
|US20040010346 *||Jul 11, 2002||Jan 15, 2004||Stewart Paul Joseph||Method of real-time collision detection between solid geometric models|
|US20040017928 *||Mar 17, 2003||Jan 29, 2004||John Herzig||Automatic compliance-testing system for desktop designed consumer packaging|
|US20040107080 *||Mar 1, 2002||Jun 3, 2004||Nikolaj Deichmann||Method for modelling customised earpieces|
|US20040119717 *||Jul 25, 2003||Jun 24, 2004||Fujitsu Limited||Animation creating/editing apparatus|
|US20040153186 *||Jan 31, 2003||Aug 5, 2004||Pravin Khurana||Horizontally-structured modeling for analysis|
|US20040153200 *||Jan 31, 2003||Aug 5, 2004||Landers Diane M.||Horizontally structured manufacturing process modeling: exterior linked representational embodiment|
|US20040153201 *||Jan 31, 2003||Aug 5, 2004||Landers Diane M.||Horizontally structured CAD/CAM coordinate system for manufacturing design|
|US20040153202 *||Jan 31, 2003||Aug 5, 2004||Landers Diane M.||Horizontally structured manufacturing process modeling: across file feature operability|
|US20040153296 *||Jan 31, 2003||Aug 5, 2004||Landers Diane M.||Horizontally structured CAD/CAM coordinate system|
|US20040153824 *||Jul 23, 2003||Aug 5, 2004||Venkat Devarajan||System and method for creating and updating a three-dimensional model and creating a related neutral file format|
|US20040194040 *||Mar 24, 2003||Sep 30, 2004||Texas Instruments Incorporated||Method of cross-mapping integrated circuit design formats|
|US20040199367 *||Apr 20, 2004||Oct 7, 2004||Koichi Kondo||Apparatus and method for obtaining shape data of analytic surface approximate expression|
|US20050033559 *||Nov 21, 2002||Feb 10, 2005||Stefan Wander||Construction method and cad system|
|US20050038642 *||Jul 23, 2004||Feb 17, 2005||Jean-Francois Rameau||Part design system using a context-free grammar|
|US20050137731 *||Dec 19, 2003||Jun 23, 2005||Albert Haag||Versioning of elements in a configuration model|
|US20050177355 *||May 20, 2003||Aug 11, 2005||Akitake Makinouchi||Cooperative work type application service system|
|US20050188335 *||Apr 19, 2005||Aug 25, 2005||Microsoft Corporation||Texture generating apparatus for dynamic interference checking|
|US20050209834 *||Mar 22, 2004||Sep 22, 2005||Landers Diane M||Horizontally structured CAD/CAM modeling-vertical to horizontal conversion|
|US20060125814 *||Feb 1, 2006||Jun 15, 2006||Minolta Co., Ltd.||Electronic apparatus|
|US20060125822 *||Feb 2, 2006||Jun 15, 2006||Alias Systems Corp.||Volume management system for volumetric displays|
|US20060136842 *||Aug 10, 2005||Jun 22, 2006||Bernard Charles||Method and computer system for interacting with a database|
|US20060143202 *||Nov 27, 2002||Jun 29, 2006||Parker Eric G||Efficient data structure|
|US20060164413 *||Jan 23, 2006||Jul 27, 2006||David Colbert||Computer-implemented geometric shape interaction process|
|US20060184343 *||Jan 25, 2006||Aug 17, 2006||Ricardo Chin||Aware and active features for computer-aided design systems|
|US20060250418 *||Dec 6, 2005||Nov 9, 2006||Dassault Corporation||Collaborative design|
|US20060256110 *||May 5, 2006||Nov 16, 2006||Yasuhiro Okuno||Virtual reality presentation apparatus, virtual reality presentation method, program, image processing method, image processing apparatus, information processing method, and information processing apparatus|
|US20060261043 *||Feb 27, 2006||Nov 23, 2006||Credence Systems Corporation||Apparatus and method for circuit operation definition|
|US20060267980 *||May 30, 2006||Nov 30, 2006||Fujitsu Limited||CAD method, CAD system and program storage medium storing CAD program thereof|
|US20060277528 *||Jun 3, 2005||Dec 7, 2006||International Business Machines Corporation||Constraining source code and objects for analysis tools|
|US20070011121 *||Jun 1, 2006||Jan 11, 2007||Jinbo Bi||System and method for learning rankings via convex hull separation|
|US20070063997 *||May 13, 2004||Mar 22, 2007||Ronny Scherer||Method and system for manipulating a digital representation of a three-dimensional object|
|US20070083826 *||May 13, 2004||Apr 12, 2007||Pessetto John R||Method and system for scaling the area displayed in a viewing window|
|US20070091119 *||Jan 26, 2006||Apr 26, 2007||Matthew Jezyk||Integrated massing and design CAD models|
|US20070146390 *||May 27, 2005||Jun 28, 2007||Canon Kabushiki Kaisha||Image processing method and image processing apparatus|
|US20070174026 *||Jan 25, 2006||Jul 26, 2007||Nicolas Mangon||Synchronized physical and analytical representations of a CAD model|
|US20070174027 *||Jan 26, 2006||Jul 26, 2007||Aleksey Moiseyev||Synchronized architectural and structural CAD models|
|US20070179759 *||Jan 31, 2006||Aug 2, 2007||Nicolas Mangon||Transferring load information and result information between analysis and design software|
|US20070179976 *||Jan 31, 2006||Aug 2, 2007||Arvin Scott A||Graphic interactive method to reorder sequential data values on graphic objects|
|US20070189602 *||Feb 6, 2007||Aug 16, 2007||Siemens Medical Solutions Usa, Inc.||System and Method for Multiple Instance Learning for Computer Aided Detection|
|US20070214432 *||Mar 8, 2006||Sep 13, 2007||Cheng-Hsun Ho||Quick naming system of graphics with the same property and method thereof|
|US20070219764 *||Dec 20, 2006||Sep 20, 2007||Autodesk, Inc.||Synchronized Physical and Analytical Flow System Models|
|US20070239406 *||Jan 25, 2006||Oct 11, 2007||Ricardo Chin||Aware and active features for computer-aided design systems|
|US20070263002 *||Jul 23, 2007||Nov 15, 2007||Pacific Data Images Llc||Stroke-based posing of three-dimensional models|
|US20070291028 *||Aug 29, 2006||Dec 20, 2007||Autodesk, Inc.||Fast Interactive Object Manipulation|
|US20070291032 *||Oct 26, 2006||Dec 20, 2007||Fujitsu Limited||Image process apparatus for three-dimensional model|
|US20080010262 *||Jun 12, 2007||Jan 10, 2008||Metacarta, Inc.||System and methods for providing statstically interesting geographical information based on queries to a geographic search engine|
|US20080010273 *||Jun 12, 2007||Jan 10, 2008||Metacarta, Inc.||Systems and methods for hierarchical organization and presentation of geographic search results|
|US20080027968 *||Jul 27, 2006||Jan 31, 2008||Autodesk, Inc.||Analysis Error Detection for a CAD Model|
|US20080143708 *||Dec 18, 2006||Jun 19, 2008||John Owen||System and method for auto-dimensioning boundary representation model|
|US20080228754 *||Oct 30, 2007||Sep 18, 2008||Metacarta, Inc.||Query method involving more than one corpus of documents|
|US20080238918 *||Apr 2, 2007||Oct 2, 2008||Autodesk, Inc.||View-specific representation of reinforcement|
|US20080252663 *||Apr 13, 2007||Oct 16, 2008||Autodesk, Inc.||Solving networks of geometric constraints|
|US20080294388 *||Mar 20, 2008||Nov 27, 2008||Thomas Tabeling||System and method for defining part interfaces|
|US20080297847 *||Aug 8, 2008||Dec 4, 2008||Fujitsu Limited||Work instruction sheet preparing device, method and program|
|US20080316204 *||Jun 19, 2008||Dec 25, 2008||Arnaud Deslandes||System and method of computer-aided design of a 3d object modeled by geometries|
|US20090055741 *||Jul 16, 2008||Feb 26, 2009||Sysmex Corporation||Work procedure display method and system, production process management method and system, and computer program of the same|
|US20090096714 *||Sep 29, 2008||Apr 16, 2009||Brother Kogyo Kabushiki Kaisha||Image display device|
|US20090096785 *||Dec 6, 2005||Apr 16, 2009||Manfred Gobel||Method For Producing Technical Drawings From 3D Models With At Least Two Colliding 3D Bodies|
|US20090256842 *||Apr 13, 2009||Oct 15, 2009||Bohman Joseph J||System and method for splitting faces on a solid model|
|US20090256843 *||Apr 13, 2009||Oct 15, 2009||Hiren Dedhia||System and method for active selection in a solid model|
|US20090259442 *||Apr 13, 2009||Oct 15, 2009||Mallikarjuna Gandikota||System and method for geometric editing|
|US20090319239 *||Jun 18, 2008||Dec 24, 2009||International Business Machines Corporation||Topology modeling application that handles abstract entities through the realization of conceptual objects|
|US20100013833 *||Jan 21, 2010||Mallikarjuna Gandikota||System and method for modifying features in a solid model|
|US20100030893 *||Jun 30, 2009||Feb 4, 2010||International Business Machines Corporation||Automated discovery of a topology of a distributed computing environment|
|US20100031247 *||Feb 4, 2010||International Business Machines Corporation||Simplified deployment modeling|
|US20100058331 *||Aug 28, 2008||Mar 4, 2010||International Business Machines Corporation||Automated deployment of defined topology in distributed computing environment|
|US20100070449 *||Sep 12, 2008||Mar 18, 2010||International Business Machines Corporation||Deployment pattern realization with models of computing environments|
|US20100077328 *||Sep 23, 2008||Mar 25, 2010||International Business Machines Corporation||Automatically arranging widgets of a model within a canvas using iterative region based widget relative adjustments|
|US20100092072 *||May 1, 2009||Apr 15, 2010||Zentech, Inc.||Automated generation of 3D models from 2D computer-aided design (CAD) drawings|
|US20100138793 *||Dec 2, 2008||Jun 3, 2010||Microsoft Corporation||Discrete objects for building virtual environments|
|US20100179788 *||Oct 13, 2009||Jul 15, 2010||Cct International, Inc.||System and method for hybrid solid and surface modeling for computer-aided design environments|
|US20100235148 *||May 27, 2010||Sep 16, 2010||Autodesk, Inc., a Delaware Corporation||Transferring Structural Loads and Displacements Between Analysis and Design Software|
|US20100235206 *||Sep 16, 2010||Project Frog, Inc.||Methods and Systems for Modular Buildings|
|US20100238167 *||Apr 13, 2009||Sep 23, 2010||Ricky Lynn Black||System and method for converting dimensions|
|US20100262950 *||Apr 9, 2009||Oct 14, 2010||On24, Inc.||Editing of two dimensional software consumables within a complex three dimensional spatial application and method|
|US20100274818 *||Oct 28, 2010||Dassault Systemes||Method and System for Updating Object Data With Respect to Object Specifications in a Product Life Cycle Management System|
|US20100280983 *||Apr 29, 2010||Nov 4, 2010||Samsung Electronics Co., Ltd.||Apparatus and method for predicting user's intention based on multimodal information|
|US20100306681 *||Feb 13, 2009||Dec 2, 2010||Dirtt Environmental Solutions Ltd.||Creation and modification of valid functional design layouts|
|US20110016074 *||Jul 16, 2009||Jan 20, 2011||International Business Machines||Method and system for encapsulation and re-use of models|
|US20110025688 *||Jul 28, 2009||Feb 3, 2011||Kevin Schneider||Method for validating features in a direct modeling paradigm|
|US20110040531 *||Apr 24, 2008||Feb 17, 2011||Siemens Ag||Method and System for Identification of Grouping Characteristics|
|US20110098982 *||Apr 28, 2011||Dassault Systemes||Method and system for updating a modeled object in a product lifecycle management system|
|US20130060530 *||Mar 7, 2013||Siemens Product Lifecycle Management Software Inc.||Tolerant intersections in graphical models|
|US20130229412 *||Nov 25, 2011||Sep 5, 2013||Sony Corporation||3d data analysis apparatus, 3d data analysis method, and 3d data analysis program|
|US20130297268 *||Aug 1, 2012||Nov 7, 2013||Siemens Product Lifecycle Management Software Inc.||Determining boolean region participants for a notional context given arbitrary bodies|
|US20140184592 *||Mar 15, 2013||Jul 3, 2014||Ketch Technology, Llc||Creating, editing, and querying parametric models, e.g., using nested bounding volumes|
|US20140336027 *||May 13, 2013||Nov 13, 2014||Xerox Corporation||System and method for creating customized boxes|
|CN102067133B||Apr 14, 2009||Apr 30, 2014||西门子产品生命周期管理软件公司||System and method for splitting faces on a solid model|
|CN102405463B *||Apr 29, 2010||Jul 29, 2015||三星电子株式会社||利用多模态信息的用户意图推理装置及方法|
|WO2007103306A2 *||Mar 6, 2007||Sep 13, 2007||Matthew Hartloff||Manipulatable model for communicating manufacturing issues of a custom part|
|WO2011033522A1 *||Feb 24, 2010||Mar 24, 2011||Steer Information Technologies Pvt. Ltd.||A method and system for configuring and visualizing an extruder|
|WO2012174274A2 *||Jun 14, 2012||Dec 20, 2012||Lucasfilm Entertainment Company Ltd.||Modifying an animation having a constraint|
|WO2014193426A1 *||Jul 16, 2013||Dec 4, 2014||Dirtt Environmental Solutions Inc.||Automatically incorporating third party features into a computer design schematic|
|U.S. Classification||703/1, 345/679, 345/420, 345/664|
|Cooperative Classification||G06T2210/12, G06T2219/2021, G06T2219/2016, G06T19/20, G06T17/10, G06F17/5086|
|May 4, 2000||AS||Assignment|
Owner name: WISCONSIN ALUMNI RESEARCH FOUNDATION, WISCONSIN
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:DANI, TUSHAR H.;GADH, RAJIT;REEL/FRAME:010745/0215;SIGNING DATES FROM 20000308 TO 20000322
|Feb 23, 2007||FPAY||Fee payment|
Year of fee payment: 4
|Feb 28, 2011||FPAY||Fee payment|
Year of fee payment: 8
|Feb 27, 2015||FPAY||Fee payment|
Year of fee payment: 12